INTERNATIONAL JOURNAL OF PSYCHOLOGY
Editor Laura Hernández-Guzmán Universidad Nacional Autónoma de Mexico, Indiana 260, Despacho 608 Mexico D.F. 03710 Mexico City, Mexico. Fax: (00-52-5) 563–6162 Email:
[email protected] Associate Editors/International Platform Pierre L.-J. Ritchie School of Psychology, University of Ottawa, 145 Jean-Jacques Lussier St., PO Box 450, Stn A, Ottawa, K1N 6N5, Canada Fax: (1) 613-562-5169 Email:
[email protected] Merry Bullock Science Directorate, APA, 750 First Street NE, Washington DC 20002, USA Email:
[email protected] (USA);
[email protected] (Estonia); URL: http://www.iupsys.org EDITORIAL BOARD The IUPsyS Executive Committee
M.Denis (President) Université de Paris-Sud, France G.d'Ydewalle (Past President), University of Leuven, Belgium J.J.Sánchez Sosa (Vice-President), Universidad Nacional Autonoma de Mexico, Mexico H.Zhang (Vice-President), Beijing Normal University, People's Republic of China P.L.-J.Ritchie (Secretary-General), Université d'Ottawa, Canada M.Sabourin (Treasurer), Université de Montreal, Canada M.Bullock (Deputy Secretary General), American
iii
Psychological Association, USA J.Adair, University of Manitoba, Winnipeg, Canada R.Ardila, National University of Colombia, Bogota, Colombia S.Cooper, c/o PsySSA, South Africa H.Imada, Kwansei Gakuin University, Japan Ç.Kagitçibasi, Koc University, Istanbul, Turkey I.Lunt, University of London, UK E.Nair, National University of Singapore J.B.Overmier, University of Minnesota, USA Y.H.Poortinga, Tilburg University, The Netherlands R.K.Silbereisen, Friedrich Schiller University of Jena, Germany
Consulting Editors
M.H.Bond, The Chinese University of Hong Kong, China C.Bruner, Universidad Nacional Autónoma de Mexico, Mexico I.Heszen-Niejodek, Silèsian University, Poland W.Kintsch, University of Colorado, USA J.Lautrey, Université de Paris V, France G.Masi, University of Pisa IRCCS Stella Maris, Italy M.McCabe, Deakin University, Australia E.Mpofu, University of Zimbabwe, Zimbabwe L.Pulkkinen, University of Jyväskylä, Finland J.Retschitzki, Université de Fribourg, Switzerland J.T.E.Richardson, The Open University, UK Z.Shechtman, University of Haifa, Israel M.Xiaochum, East China Normal University, China
French Translation Consulting Editor
iv
J.Emery, Canada
Aims and Scope
The International Journal of Psychology is the journal of the International Union of Psychological Science (IUPsyS) and is published under the auspices of the IUPsyS. Its purpose is to circulate, in an international framework, scientific information within and among subdisciplines of psychology and to foster the development of psychological science around the world. The journal emphasizes empirical research and theory in basic and applied psychology, especially with an international focus. It does not publish technical articles, validations of questionnaires and tests, or clinical case studies. Regular issues include two types of articles: empirical articles and review articles. Empirical articles report data from single or multiple studies in one of the major fields of scientific psychology. Review articles provide overviews of the international literature on a particular topic; authors are especially encouraged to include in their review relevant publications from regions of the world not typically cited and/or in languages other than English. Special topical issues or sections are also published two or three times a year. The International Platform for Psychologists Associate Editors: M.Bullock (USA); P.L.-J Ritchie (Canada) Many of IJP's issues include a second section, the International Platform for Psychologists, which provides an opportunity to exchange news and opinions on psychology as an academic and applied profession. This section also contains information about the IUPsyS, about major international meetings, and about the activities of the National Psychological Societies. Finally it offers an opportunity to express opinions and to discuss internationally significant psychological issues. There is also a United Nations section with the International Platform for Psychologists.
Publication. The International Journal of Psychology is published by Psychology Press Ltd, a member of theTaylor & Francis group, 27 Church Road, Hove, East Sussex, BN3 2FA, UK, on behalf of the International Union of Psychological Science. The International Journal of Psychology is now available online: Psychology Online on www.psypress.co.uk for more information (online ISSN: 1464–066X). Subscription rates to Volume 38, 2003 (6 issues) are as follows: Individuals: UK: £75.00; Individuals: Rest of world: $123.00 Institutions: UK: £321.00; Institutions: Rest of world: $529.00 Plus free Psychology: IUPsyS Global Resource CD-ROM to all subscribers International Journal of Psychology (USPS permit number 016263) is published bimonthly in February, April, June, August, October, and December. The 2003 US Institutional subscription price is $529.00. Periodicals postage paid at Champlain, NY, by US Mail Agent IMS of New York, 100 Walnut Street, Champlain, NY. US Postmaster: Please send address changes to pIJP, PO Box 1518, Champlain, NY 12919, USA. Subscription orders should be addressed to: Psychology Press, Taylor & Francis Ltd, Rankine Road, Basingstoke, Hampshire, RG24 8PR, UK; Fax: +44 (0) 1256 479438. Please give six weeks' notice of changes of address and include both old and new addresses. Permissions and fee are waived for the photocopying of isolated articles for non-profit classroom and library reserve use by instructors and educational institutions. This consent does not extend to other kinds of copying, such as copying for general distribution for advertising or for promotional purposes, for creating new collective works, or for resale. The International Journal of Psychology is covered by the following abstracting, indexing, and citation services: Assia; Biosis; Bell & Howell Learning (formerly University Microfilms); Current Contents - Social/ Behavioural Sciences (ISI); Ergonomics Abstracts; Linguistic and Language Behavior Abstracts; NISC/Family Studies Database; Psychological Abstracts/ PsychINFO; Research Alerts (ISI); Social SciSearch (ISI); Social Science Citation Index (ISI); Social Services Abstracts (formerly SOPODA); Sociological Abstracts; and UnCover. Typeset by Digital Publishing Solutions Ltd, Pune, India Printed in the United Kingdom by Henry Ling Limited, at the Dorset Press, Dorchester, DT1 1 HD. This publication has been produced with paper manufactured to strict environmental standards and with pulp derived from sustainable forests.
© 2003 International Union of Psychological Science
Prospective memory research: Why is it relevant? Matthias Kliegel and Mike Martin University of Zurich, Switzerland
“Which were the three most important memory problems you had in the last week?” We put this question to all participants who attended memory training classes in an adult education institution in a mid-sized German town over the last 2 years. Of those interviewed, 62% indicated that at least one of these problems concerned the delayed execution of an intended action, e.g., forgetting to give someone a call. This type of memory task has been labeled prospective memory and interest in this rather new field of cognitive psychology is growing (Brandimonte, Einstein, & McDaniel, 1996). There are at least three reasons why research in prospective remembering is highly relevant. (1) As signified by the finding reported above, prospective memory is of great relevance for everyday life. Various studies have reported that 50–80% of all everyday memory problems are, at least in part, prospective memory problems (e.g., Crovitz & Daniel, 1984; Terry, 1988). (2) Prospective memory is of enormous clinical relevance. In a current study investigating patients attending the Heidelberg Memory Clinic, to date about 40% of the patients have reported prospective memory problems as their main symptoms. Moreover, in the last years, several studies addressed the question of prospective memory problems in neuropsychological patients (e.g., Fortin, Godbout, & Braun, 2002). Recent efforts in this context concern the mechanisms of these impairments (see Kopp & Thöne-Otto, this issue) as well as possible strategies of rehabilitation (see Thöne-Otto & Walther, this issue). (3) Prospective memory research is of tremendous theoretical relevance. Aiming to disentangle prospective memory from the traditional topic of memory research, i.e., the memory for previously learned information or retrospective memory, the question of “To what extent is prospective memory similar to and to what extent is it different from retrospective memory?” (Guynn, McDaniel, & Einstein, 2001, p. 25) still remains. However, in addressing this issue, concepts of cognitive functioning in general are being advanced (see Guynn, this issue). In addition, one major focus of prospective memory research has been the life-span development of prospective remembering. Investigating age effects and possible underlying mechanisms, the influence of executive functioning (see Martin, Kliegel, & McDaniel, this issue), the specific nature of memory for intentions
2 KLIEGEL AND MARTIN
(see Freeman & Ellis, this issue), and the role of motivational aspects (see Kliegel, Martin, & Moor, this issue) are currently being examined. In sum, the present special issue aims to present a variety of examples of current approaches in the field of prospective memory research. We hope that we will stimulate further research in these and other aspects of the question: Why do we forget to execute intended actions every day? REFERENCES Brandimonte, M., Einstein, G.O., & McDaniel, M.A. (Eds.) (1996). Prospective memory: Theory and applications . Mahwah, NJ: Lawrence Erlbaum Associates Inc. Crovitz, H.F., & Daniel, W.F. (1984). Measurements of everyday memory: Toward the prevention of forgetting. Bulletin of the Psychonomic Society , 22 , 413–414 . Fortin, S., Godbout, L., & Braun, C.M.J. (2002). Strategic sequence planning and prospective memory impairments in frontally lessoned head trauma patients performing activities of daily living. Brain and Cognition , 48 , 361–365 . Freeman, J.E., & Ellis, J.A. (this issue). The intention-superiority effect for naturally occurring activities: The role of intention accessibility in everyday prospective remembering in young and older adults. International Journal of Psychology , 38 , 215–228 Guynn, M.J. (this issue). A two-process model of strategic monitoring in event-based prospective memory: Activation/retrieval mode and checking. International Journal of Psychology , 38 , 245–256 . Guynn, M.J., McDaniel. M.A., & Einstein, G.O. (2001). Remembering to perform actions: A different type of memory? In H.D.Zimmer, R.L.Cohen, M.J.Guynn, J.Engelkamp, R.Kormi-Nouri, & M.A.Foley (Eds.), Memory for action: A distinct form of episodic memory ? ( pp. 25–48 ). New York: Oxford University Press. Kliegel, M., Martin, M., & Moor, C. (this issue). Complex prospective memory and aging: The influence of motivational incentives. International Journal of Psychology , 38 , 207–214 . Kopp, U.A., & Thöne-Otto, A.I.T. (this issue). Disentangling executive functions and memory processes in event-based prospective remembering after brain damage: A neuropsychological study. International Journal of Psychology , 38 , 229–235 . Martin, M., Kliegel, M., & McDaniel, M.A. (this issue). The involvement of executive functions in prospective memory performance of adults. International Journal of Psychology , 38 , 195–206 . Terry, W.S. (1988). Everyday forgetting: Data from a diary study. Psychological Reports , 62 , 299–303 . Thöne-Otto, A.I.T., & Walther, K. (this issue). How to design an electronic memory aid for brain injured patients: Considerations on the basis of a model of prospective memory. International Journal of Psychology , 38 , 236–244 .
The involvement of executive functions in prospective memory performance of adults Mike Martin and Matthias Kliegel University of Zurich, Switzerland Mark A. McDaniel University of New Mexico, Albuquerque, USA
The present study examines the relationship between prospective memory performance and executive functioning. The four phases of the prospective memory process—intention formation, intention retention, reinstantiation of the intention, and intention execution— are assumed to require different amounts of executive processing, most of which is demanded in the phases of intention formation and intention execution. At present, though, it is still unclear whether, and to what extent, prefrontal executive systems are involved in different kinds of prospective memory tasks, as some findings suggest that prospective memory might rather rely on nonstrategic processes that are unlikely to depend on prefrontal executive systems. Therefore, this study focuses on the following questions: (1) to what degree does executive functioning predict prospective memory performance in different standard prospective memory tasks and, furthermore, are certain executive measures better predictors than others; (2) are age-related effects in different prospective memory measures due to individual differences in executive functioning; and (3) do age-related differences in prospective memory exist that are not explained by individual differences in executive functioning. In a sample of 80 adults (20–80 years), we applied four instruments to measure prospective memory: a traditional single-task paradigm, two more complex tasks—one timebased and one event-based—and a highly complex multi-task paradigm. We further assessed a broadly defined construct of executive functioning using several standard neuropsychological tests. Results showed that executive functioning did not predict performance in the simple single-task paradigm. However, executive functioning, but not age, predicted performance in the two more complex standard tests of prospective remembering, and both executive functioning and age predicted performance in the most complex paradigm. In sum, the obtained data underline the assumption that frontal/ executive functions are related to
4 MARTIN, KLIEGEL, MCDANIEL
prospective memory performance across a range of prospective paradigms. It also seems clear that age differences in prospective memory performance partially depend on age-related individual differences in frontal/executive functions. La présente étude examine la relation entre la performance de la mémoire prospective et le fonctionnement exécutif. Il est présumé que les quatre phases du processus de la mémoire prospective (la formation de l’intention, la rétention de l’intention, la représentation d’une abstraction par l’utilisation d’un exemple concret de l’intention et l’exécution de l’intention) requièrent des traitements exécutifs d’ampleur variable, la plupart étant plus exigeants dans les phases de la formation de l’intention et de l’exécution de l’intention. Cependant, actuellement, la façon et l’ampleur avec lesquelles les systèmes préfrontaux exécutifs sont impliqués dans les différents types de tâches de la mémoire prospective ne sont pas claires étant donné que les résultats d’étude suggèrent que la mémoire prospective ferait plutôt appel a des processus non stratégiques étant probablement indépendants des systèmes préfrontaux exécutifs. La présente étude vise a examiner les questions suivantes: (1) à quel degré le fonctionnement exécutif prédit la performance de la mémoire prospective dans différentes tâches prospectives standards et, par ailleurs, est-ce que certaines mesures executives sont de meilleurs prédicteurs que d’autres?; (2) est-ce que les effets de l’âge dans les différentes mesures de mémoire prospective sont dus aux differences individuelles dans le fonctionnement exécutif?; (3) est-ce qu’il existe des differences reliées a l’âge dans la mémoire
© 2003 International Union of Psychological Science http:// DOI: 10.1080/ www.tandf.co.uk/ 00207590244000205 journals/pp/ 00207594.html
Requests for reprints should be addressed to Mike Martin, Institute of Psychology, Department of Gerontopsychology, University of Zurich, Schaffhauserstr. 15, CH-8006 Zurich, Switzerland (E-mail:
[email protected]). Part of the preparation of this manuscript was supported by DFG grant MA 1896/4–1. Appreciation is expressed to Sonja Barth for experimental assistance, and to Caroline Moor for helpful comments on a prior version of the manuscript.
PROSPECTIVE MEMORY PERFORMANCE
prospective qui ne sont pas expliquées par les differences individuelles dans le fonctionnement exécutif? Nous avons administré, auprès d’un échantillon de 80 adultes (20–80 ans), quatre instruments pour mesurer la mémoire prospective:un paradigme traditionnel de tâche simple, deux tâches plus complexes—une basée sur le temps et l’autre basée sur lesévénements—et un paradigme multi-tâches très complexe. Nous avons également évalué un construit de fonctionnementexécutif largement défini en utilisant plusieurs tests neuropsychologiques standards. Les résultats ont soulevé que lefonctionnement exécutif ne prédit pas la performance dans les paradigmes de tâche simple. Cependant, le fonctionnementexécutif prédit la performance dans les deux tests standards plus complexes de rappel prospectif, ce qui n’est pas le cas del’âge. Enfin, a la fois le fonctionnement exécutif et l’âge prédisent la performance dans le paradigme le plus complexe. Ensomme, les données obtenues soutiennent la prediction que les fonctions frontales/exécutives sont reliées a la performancedans la mémoire prospective a travers une gamme de paradigmes prospectifs. Aussi, il apparaît clairement que les differencesd’âge dans la performance de la mémoire prospective dependent partiellement des differences individuelles reliées a l’âgedans les fonctions frontales/exécutives. El presente estudio examina la relación entre el desempeño de la memoria prospectiva y el funcionamiento ejecutivo. Se supone que las cuatro fases del proceso de memoria prospectiva -la formación de la intención, la retención de la intención, la representación de una abstracción por medio de una instancia concreta de la intención y la ejecución de la intención—requieren diferentes cantidades de procesamiento ejecutivo, del cual hay más demanda en las fases de formación de la intención y ejecución de la intención. Al presente, sin embargo, no queda claro si los sistemas ejecutivos prefrontales están involucrados en diferentes tipos de tareas de memoria prospectiva, y en que medida, ya que algunos hallazgos sugieren que la memoria prospectiva depende más bien de procesos no estratégicos que es poco probable que dependan de los sistemas ejecutivos prefrontales. Por consiguiente, el presente estudio se centra en las siguientes preguntas: (1) ¿en que medida el funcionamiento ejecutivo predice el desempeño de la memoria prospectiva en diferentes tareas tipificadas de memoria prospectiva y, es más, si ciertas medidas predicen mejor que otras?; (2) ¿se deben los efectos relativos a la edad en diferentes medidas de memoria prospectiva, a diferencias individuales en el funcionamiento ejecutivo? y (3) ¿existen diferencias relativas a la edad en la memoria prospectiva que no pueden explicarse en términos de diferencias individuales en el funcionamiento ejecutivo? Una
5
6 MARTIN, KLIEGEL, MCDANIEL
muestra de 80 adultos (20–80 años) respondieron a cuatro instrumentos que median memoria prospectiva: un paradigma tradicional de una sola tarea, dos tareas más complejas -una basada en tiempo y la otra en el evento, y un paradigma multi-tarea muy complejo. Se evaluó además un constructo de funcionamiento ejecutivo definido ampliamente, por medio del uso de varias pruebas neuropsicológicas estándar. Los resultados muestran que el funcionamiento ejecutivo no predice el desempeño en el paradigma de una sola tarea simple. Sin embargo, el funcionamiento ejecutivo, más no la edad, predice el desempeño en dos o más pruebas estándar de recuerdo prospectivo, y tanto el funcionamiento ejecutivo como la edad predicen el desempeño en el paradigma más complejo. En resumen, los datos obtenidos subrayan el supuesto de que las funciones frontales/ejecutivas se relacionan con el desempeño de la memoria prospectiva a lo largo de un espectro de paradigmas prospectivos. También parece ser que las diferencias de edad en el desempeño de la memoria prospectiva dependen parcialmente de diferencias individuales relativas a la edad en las funciones frontales/ ejecutivas. Remembering to perform an intended action at a particular point in the future, i.e., prospective memory, is essential in the everyday life of older adults, because self-initiated acting upon earlier formed intentions is at the core of their independent living. Without it or with low levels of performance, appointments (Kvavilashvili, 1987; Ma. Martin, 1986; West, 1988), medication (Park & Kidder, 1996), or, more generally, chances to act are likely to be missed (Ellis, 1996; Maylor, 1993). However, research focusing on prospective memory in old age is in its infancy (cf. Roediger, 1996), with several fundamental issues just beginning to receive theoretical and empirical attention. One such issue concerns the neuropsychological systems and processes that support prospective remembering. An initial proposal is that prospective memory performance depends on the prefrontal systems and the integrity of the executive functions that these systems subserve (Bisiacchi, 1996; Burgess, Veitch, de Lacy Costello, & Shallice, 2000; Glisky, 1996; McDaniel, Glisky, Rubin, Guynn, & Routhieaux, 1999; see also Stuss & Benson, 1987). Frontally mediated executive functions are believed to include planning, interruption and inhibition of responses, monitoring of environmental events, and flexible initiation of responses to those events (Shimamura, Janowsky, & Squire, 1991). The prospective memory process consists of the four phases of intention formation, intention retention, reinstantiation of the intention, and execution of the reinstantiated intention (Ellis, 1996; Kliegel, Martin, McDaniel, & Einstein, in press; Kvavilashvili & Ellis, 1996). The theoretical assumption is that frontal functions are little involved in the retention phase, and intimately involved in performance of the
PROSPECTIVE MEMORY PERFORMANCE
7
intention formation and intention execution phase of prospective memory tasks. Older adults’ prospective performance in tasks that emphasize the phases theoretically making the strongest demands on the executive functions should be highly related to executive function measures (cf. West, 1996). However, there is virtually no empirical work that has attempted to evaluate this assumption. Further, the limited work that is available shows somewhat mixed findings. Some studies report relations between frontal processes and prospective remembering in adults (Bisiacchi, 2000; Burgess, 2000a; Burgess et al., 2000; Kopp & Thöne, 2000; McDaniel et al., 1999), whereas others do not necessarily implicate frontal processes (Bisiacchi, 1996; Cockburn, Keene, & Hope, 2000; Mi.Martin, Kliegel, McDaniel, & Einstein, 2000). These mixed patterns suggest several possibilities. One clear possibility is that prospective memory in adulthood involves a range of processes that depend on the particular instantiation of the prospective memory task (cf. Einstein & McDaniel, 1990; McDaniel & Einstein, 2000). While some prospective memory tasks appear to require prefrontal executive processes of planning, monitoring, or flexibility in response preparation, others seem to rely on relatively nonstrategic processes, i.e., processes that might not depend on prefrontal executive systems (Guynn, McDaniel, & Einstein, 2001; Kliegel, Martin, McDaniel, & Einstein, 2001; Kliegel, McDaniel, & Einstein, 2000; McDaniel & Einstein, 2000; McDaniel, Robinson-Riegler, & Einstein, 1998). The amount of required executive processing is likely to depend on the degree to which the dependent prospective measures focus on the intention formation and/or execution phases versus the retention phase. The more relative weight is given to intention formation and intention execution, the stronger should be the relation to executive functioning. To inform this possibility, the first goal of the present study was to investigate the contribution of prefrontal executive processes to performance in four standard prospective memory tasks. As a consequence, we test if differences in executive functioning are related to differences in prospective memory performance in young and old adults. Unlike earlier studies with only old adults (e.g., McDaniel et al., 1999), we compare prospective memory performance and its relation to executive functioning between young and old adults in four different prospective memory tasks. The selected prospective memory tasks differentially represent several major components on which such tasks can vary. One critical component is whether the task is time-based or event-based (Einstein & McDaniel, 1990, 1996). Timebased prospective memory involves remembering to perform the intended action at a certain time or after a certain amount of time has elapsed (e.g., remembering to take cookies out of the oven in 10 minutes). Presumably, time-based tasks (for which no external reminders are implemented) are dependent on self-initiated monitoring in the intention execution phase of the prospective task, an assumption consistent with age-related deficits in monitoring patterns and subsequent prospective memory performance (Einstein, McDaniel, Richardson,
8 MARTIN, KLIEGEL, MCDANIEL
Guynn, & Cunfer, 1995) and with increased time-based performance as a function of the importance of the task (Kliegel et al., 2001). Based on the idea that prefrontal functions control monitoring (Shallice & Burgess, 1991; Shimamura et al., 1991), prefrontal systems would be expected to be significantly associated with time-based prospective memory. No published study has yet examined this prediction 1 , and for this reason one of our prospective tasks was a standard laboratory time-based task (Einstein, Smith, McDaniel, & Shaw, 1997; Kliegel et al., 2001). Event-based prospective memory involves remembering to perform the intended action when some external event occurs (such as remembering to give a message to a colleague when passing her office). Thus, in contrast to time-based prospective memory, event-based tasks have external cues that can stimulate retrieval of the intended action. In a sense, the event-based task is similar to cued recall (McDaniel & Einstein, 1993), and thereby may attenuate the degree of selfinitiated retrieval or monitoring required to remember the intended action at the appropriate moment in the intention execution phase of the prospective task (Einstein & McDaniel, 1990, 1996). For instance, in an event-based prospective memory task introduced by Einstein et al. (1997), participants were instructed to press a particular function key on a computer keyboard when a specified target item appeared in a word-rating task. Age-related declines on the task were modest (Einstein et al., 1997, Experiment 1), and varying the importance of the task did not change performance levels (Kliegel et al., 2001). Both findings suggest that relatively few strategic retrieval processes support this event-based task. To evaluate this idea, our second and third prospective memory tasks were two standard event-based prospective memory tasks. As a standard laboratory task, the event-based task used in Kliegel et al. (2001; cf. Einstein et al., 1997) was administered. In addition, as a standard clinical event-based prospective memory task, the Remembering-a-Belonging subtest from the Rivermead Behavioral Memory Test (RBMT; Wilson, Cockburn, & Baddeley, 1985; for applications: e.g., Cockburn & Smith, 1988, 1991, 1994; Evans & Wilson, 1992; Hon, Huppert, Holland, & Watson, 1998; Huppert & Beardsall, 1993; Wisemann, Ratcliff, Chase, Laporte, Robertson, & Colantonio, 2000) was applied. In this task, participants are required to demand the return of an item at the end of the session before leaving the room. Because there is only a single event-based action that has to be remembered, we reasoned that performance on the RBMT should not rely on strategic executive control processes in the intention execution phase of the task. Our fourth prospective memory task was a complex prospective memory task patterned on that used by Kliegel et al. (2000; see also Burgess, 1996; Burgess et al., 2000). We selected this task because (1) participants are required to execute a
1
Bisiacchi (1996) tested time-based prospective memory but did not separate it from event-based prospective memory in relating performance to neuropsychological variables.
PROSPECTIVE MEMORY PERFORMANCE
9
set of six intended actions rather than a single action after the retention (i.e., delay phase); (2) because the appropriate delayed execution of the set of actions involves event-based and time-based characteristics; and (3) because planning is explicitly required in the intention formation phase of this task. We reasoned that these complex demands in both the intention formation and the intention execution phases should require the most robust degree of prefrontal executive involvement, if such involvement indeed is central to prospective memory performance in the respective task phases. Another possibility for the mixed patterns from the initial investigations of the importance of prefrontal executive processes in prospective remembering is that executive functioning may have been inadequately captured by the measures used. As noted above, prefrontal systems are thought to support a wide array of executive functions, and it is unlikely that one measure will adequately capture the integrity of these executive functions. Accordingly, we used a broadly defined construct of executive functioning based on a set of three executive functioning tasks from clinical and experimental literature, rather than restricting focus to just one specific measure (for a similar procedure see McDaniel et al., 1999). Further, we used assessments that are thought to index particular components of frontal function, such as planning, inhibition, monitoring ongoing activity, and cognitive and response flexibility, that are theoretically related to prospective memory performance as sketched above. A second goal was to explore a corollary hypothesis to the idea that prefrontal processes are intimately involved in prospective memory performance. Based on the idea that frontal functioning may show preferential decline with age (e.g., Schretlen et al., 2000; Wecker, Kramer, Wisniewski, Delis, & Kaplan, 2000; West, 1996), we suggest that age-related decline in prospective memory will be significantly associated with decline in frontal/executive functioning. Others have made similar proposals (e.g., Glisky, 1996) and there is some preliminary data to support this hypothesis (Kliegel et al., 2000) but, to date, this hypothesis has not been specifically tested. The idea is that the age-related differences found in prospective memory will be eliminated once the variance in performance due to frontal/executive functioning is taken into account. In this study, by using several prospective memory tasks on which either significant age-related differences have been established (time-based, e.g., Einstein et al., 1995; Park, Hertzog, Kidder, Morrell, & Mayhorn, 1997; complex prospective memory, Kliegel et al., 2000,2001, 2002) or tendencies for age-related differences are evident (event-based, cf. Cherry & LeCompte, 1999; Einstein et al., 1997; Huppert, Johnson, & Nickson, 2000; Maylor, 1996; Park et al., 1997; but cf. Kliegel et al., 2000, for different findings concerning the RBMT), we are able to test a more complete version of this hypothesis. Alternatively, it might be that there is some age-related component of prospective memory performance that is not mediated by frontal/executive processes. That is, even after factoring out effects of frontal/executive processes, age deficits may remain. If so, then we will have support for the assumption that prospective memory signals age-related
10 MARTIN, KLIEGEL, MCDANIEL
decline that is differentiated from typical markers of age-related frontal decline (cf. Wecker et al., 2000). Overall, we expect the weakest effect of executive functioning differences on the RBMT performance, a medium effect of executive functioning differences on the standard experimental event- and time-based paradigm performances, and a strong effect of executive functioning differences on the performance in the complex prospective paradigm. In addition, when groups of young, old/high executive functioning, and old/low executive functioning groups are compared, we expect no effects of age and executive functioning differences in the RBMT performance, effects of executive functioning differences on the performances in the experimental event- and time-based paradigm performances, and effects of executive functioning differences and age on the performance in the complex prospective paradigm. The difference in the relations can be explained by the differential emphases that these prospective memory paradigms lay on the intention formation and the intention execution phases. Whereas the RBMT focuses on the retention and reinstantiation component in a least demanding environment, the standard experimental paradigms focus on the intention execution in a somewhat demanding task environment. Finally, the complex prospective paradigm maximizes the interindividual differences in the intention formation and the intention execution phases in a demanding task environment, thus theoretically requiring the largest involvement of executive functions. METHOD Participants The 80 participants in this study, including 40 young and 40 old adults, completed the battery of tests in an average time of 120 minutes, distributed over two testing sessions an average of 1 week apart. The young participants were students of the University of Mainz who volunteered. The old participants were community dwelling volunteers. As Table 1 indicates, despite small differences between the groups with respect to gender, health, or years of formal schooling, which are to be expected when recruiting samples of young and old adults, the two groups seem largely comparable. Instruments Prospective memory measures The complex or multitask prospective memory paradigm (MTPM; Kliegel et al., 2000, Kliegel, Martin, McDaniel, & Einstein, 2002a; Martin et al., 2000) is a modified six-elements task after Shallice and Burgess (1991). The task consists
PROSPECTIVE MEMORY PERFORMANCE
11
of three phases: an introduction phase, in which the participants are required to develop a plan for executing the prospective task, a delay phase, in which individual difference variables can be assessed, and a performance phase, in which the multitask prospective memory paradigm had to be executed. In the introduction phase, participants were told that at a certain point—when they would have to fill out a personal information questionnaire—they would be required to execute a set of six tasks. The participants were informed that this would take place during the second part of the experiment, after a short TABLE 1 Demographic characteristics of young and old adults
Age
Young adults (N=40)
Old adults (N=40)
M=24.8 SD=2.0 Range=22–31
M=69.3 SD=5.6 Range=60–80
Gender Female Male Years of education
N=21 N=19 M=13.2 SD=0.9 a M=3.9 Health SD=0.9 aRated on a scale from 1 (poor) to 5 (excellent).
N=30 N=10 M=12.2 SD=2.1 M=3.4 SD=0.7
break and some other tasks. Using example sheets, the tasks and the rules of our modified six-elements task were explained to the participant. Specifically, the participants were asked to remember to carry out six subtasks in a 6-minute period of time. The six subtasks were divided into two similar sets (sets A and B) of three (word finding, solving arithmetic problems, and writing down the names of pictures). We designed each subtask so that it would need more than 1 minute to complete. After the subtasks were explained, participants were told where the material for these subtasks was stored (in the second drawer of the participant’ s table, divided in three file-folders according to the task type) and that there would be a few rules to follow. The rules explicated that earlier problems would be given more points than later ones in each subtask, that each of the six subtasks would be given equal weight, and that doing two subtasks (A) and (B) of the same type one after the other was not allowed. When they were aware of the rules, i.e., were able to recall them perfectly, the participants were told that they would have to remember to work on all six tasks, and that they would have to open the drawer and initiate these tasks by themselves after answering the question about their date of birth in the “participant information” questionnaire that had previously been explained. Finally, the participants were asked to verbally
12 MARTIN, KLIEGEL, MCDANIEL
develop a plan for the multitask prospective memory paradigm. After a filled delay (in which inhibition and event-based prospective memory were assessed), the experimenter made sure that all participants reinstantiated the original execution intention, i.e., participants not initiating the multitask prospective memory paradigm by themselves were prompted by the experimenter. This way, the focus of the prospective memory task is clearly on the intention planning and the execution phases, whereas retention and reinstantiation are kept stable between participants. After having worked on the six elements task for 6 minutes, the participants filled out the rest of their “participant information” questionnaire. We also interviewed all participants about how well they remembered the instructions and their original plans. Finally, participants were debriefed by the experimenter. The dependent variables are the number of subtasks started (MTPM/ performance) and the percentage of recalled executable actions of the original plan from the planning phase of the MTPM as an indicator of retrospective memory (note that for experimental purposes we tried to minimize age differences in this measure). From experimental psychology literature, we selected two paradigms. One involves giving the person a specific cue that will be presented during their ongoing activity, thus putting a strong emphasis on the execution phase of the task. For this laboratory version of an event-based prospective memory task, all participants weregiven a slightly modified standard prospective memoryparadigm introduced by Einstein et al. (1997). Theongoing task was a computerized word rating task, inwhich 26 words (e.g., house, phone, love, war, etc.) hadto be rated on four dimensions (concreteness, familiarity,pleasantness, and seriousness). On each trial one wordwas presented with one dimension and a rating scalefor 5 seconds on the computer screen. The rating hadto be done by pressing the corresponding number keyon the computer keyboard. All 26 words were presentedfour times in the same order with changing ratingdimensions. The presentation order of the ratingdimensions was randomized but one dimension wasnever presented twice in a row. Hence, 104 trials werepresented to every participant. The prospective memorytask was to press a target key whenever the Germanword “Gespräch” (i.e., conversation) appeared on thescreen as a word to be rated. The target word appearedevery 2 minutes, and the task lasted 8 minutes and 40 seconds. Another experimental prospective memory task involves requiring the subject to respond after a particular time interval, thus also emphasizing the execution phase of the prospective task. For this laboratory version of a time-based prospective memory task, the materials were identical to those used in the eventbased version with two exceptions: The prospective memory task was to press a target key on the computer keyboard every 2 minutes after having started. The response was scored as correct if it occurred within a time window of plus or minus 5 seconds. Again, all instructions and materials were presented on the computer screen, and the task lasted 8 minutes and 40 seconds. In addition,
PROSPECTIVE MEMORY PERFORMANCE
13
participants were told that they could monitor the time by pressing a yellow key, when a time counter clock would appear for 2 seconds. Our fourth standard measure of prospective memory was selected from clinical/assessment literature. The task was the Rivermead Behavioral Memory Test (RBMT; Wilson et al., 1985), from which we selected the Remembering-aBelonging subtest. In this test, the experimenter requests a personal belonging from the participant at the beginning of the experimental session and instructs the participant to demand the return of the item at the end of the session before leaving the room, without any additional reminding on the part of the experimenter. The dependent variable is the correct request to return the belonging after a delay of approximately 45 minutes. In this test, there is little emphasis on the intention planning and intention execution phases, but, instead, the retention and reinstantiation of the intention are emphasized. Executive functioning measures In order to measure executive functions, a number of instruments have been developed and are currently in clinical and research use. We selected three prominent measures, trying to capture the essential aspects of executive functioning. The calculated factor score derived from the raw scores can be used for individual analyses and to split the groups into young, old/high executive functioning, and old/low executive functioning. The Wisconsin Card Sorting Test (WCST; Heaton, Chelune, Talley, Kay, & Curtiss, 1993) is a computerized test to measure the ability to shift cognitive strategies or sets in response to changing environmental contingencies (e.g., Greve, Brooks, Crouch, & Williams, 1997). In the test, participants have to sort a deck of 64 cards to match 1 of 4 stimulus cards. The computer tells the participant if the answer is right or wrong. Once the participant has made 10 consecutive “correct” matches to the initial sorting principle (i.e., colour), the sorting principle is changed to form or number without a warning, requiring the participant to use the feedback to develop a new sorting rule or, in other words, to shift sets. The dependent variable is the number of categories participants have completed after having sorted 64 cards. We included a colour-word version of the Stroop task (cf. Stroop, 1935; for applications cf. Houx, Jolles, & Vreeling, 1993) in order to measure inhibition (e.g., Dempster, 1992; Spieler, Balota, & Faust, 1996). In this task, the word stimuli consisted of four colour names (red, blue, green, and yellow) written in black ink (trial 1: read the words as fast as you can), the four colour names printed in colour bars (trial 2: name the colours as fast as you can), and the four colour names printed in mismatching colours (trial 3: name the colours the words are written in as fast as you can). Each Stroop condition began with practice of the top line consisting of 5 items, followed by timed performance on the 20 test items (consisting of 5 rows of 4 items each). The dependent variable is the time difference between the baseline (colour bars) and the interference condition.
14 MARTIN, KLIEGEL, MCDANIEL
The Tower of London (TOL) is a planning task in which difficulty can be varied by using different end states that have to be achieved starting from a given starting state (e.g., Morris, Evenden, Sahakian, & Robbins, 1987; Ward & Allport, 1997). The participants’ task is to make the fewest possible moves of different-sized disks to reach the end state. The dependent variable is the difference between the minimum number of moves and the number of moves the participants made. Results from the principal components analysis with z-scores of all three executive measures suggested a single executive factor representing the common variance of all three executive tests and explaining 57.6% of the variance. Component loadings were in the expected directions and ranged between .70 to . 80. Using the standardized factor score (for a similar procedure see McDaniel et al., 1999), a three-way median split between all participants in the study resulted in a young adult group and two old adult groups. The young adults had the lowest mean score, i.e., the highest performance (M=–0.74, SD=0.39), the old adults with high executive performance were close to the mean of the distribution (M=– 0.02, SD=0.42), and the old adults with low executive performance showed the highest mean score, i.e., the lowest performance (M=1.36, SD=0.72). All group differences were significant. Procedure In the course of the first session, the participants were given the Rivermead Behavioral Memory Test and the time-based prospective memory task. After a short break of 5 minutes the TOL as well as the WCST were administered. In the course of the second session, participants were first instructed in the complex or multitask pro spective paradigm, then had to develop their plan for this task, and were next tested on the Stroop task as well as on the event-based prospective memory task. After a short break of 5 minutes, participants worked on a distractor task. After completing the distractor task, they were given the “participant information” questionnaire as the cue for the multitask prospective memory paradigm. After working for 6 minutes on the complex prospective memory task, participants finished the participant-information questionnaire, responded to the questions asking what respondents remembered from their original plans, and were debriefed by the experimenter. RESULTS The first goal of our study was to assess and represent a wide range of typical standard prospective memory tasks as well as of frontal executive processes. Table 2 summarizes the results in these prospective and executive measures. The results indicate increasing age differences the more strongly frontal/ executive functions are hypothesized to be involved in performing the prospective task. There are age differences for all executive measures, for the
PROSPECTIVE MEMORY PERFORMANCE
15
laboratory event- and time-based tasks, and marked age effects for the MTPM/ performance measure. There was no age effect in the single action event-based task from the RBMT. In addition, we intended to explore the extent to which individual differences in executive functioning are associated with individual differences in prospective memory performance. Therefore, we conducted a correlation analysis for young and old adults with the prospective memory measures and the executive functioning factor score. The results can be found in Table 3. The results indicate that only in older adults were the laboratory event- and time-based prospective memory tasks as well as the complex prospective memory measure significantly correlated with executive functioning. The RBMT measure was not correlated to the executive factor score in either age group. A second goal was to investigate if individual differences rather than age per se explain most of the age-related variance in prospective memory tasks. This would be suggested if, after controlling for individual differences in nonexecutive functions, age-related differences in prospective memory are eliminated once the variance in performance due to executive functioning is taken into account. Alternatively, there might be some age-related component of TABLE 2 Age differences in executive functioning, and prospective memory measures Young adults (N=40) Executive measures WCST 5.8 (0.9) Stroop 10.3 (4.4) Tower of London 3.5 (3.4) Prospective memory measures RBMT (% correct) 62.5 Lab. event-based 3.79 (0.5) Lab. time-based 3.89 (0.4) MTPM/performance 5.50 (0.8) *p<.05; **p<.01; ***p<.001.
Old adults (N=40)
F-value
η2
4.0 (1.8) 26.4 (14.0) 9.4 (5.7)
31.98*** 38.12*** 29.17***
.296 .328 .280
82.5 3.15 (1.5) 2.97 (2.1) 2.83 (1.3)
0.48 6.35* 6.32* 127.18***
.007 .077 .081 .620
TABLE 3 Correlations between prospective memory measures and executive functioning for young and old adults Prospective memory task
Executive functioning (factor score)
RBMT Event-based Young/Old Young/Old .17/.20 .14/.40*
Time-based Young/Old .32+/.32*
MTPM/Performance Young/Old .19/.41**
16 MARTIN, KLIEGEL, MCDANIEL
Prospective memory task +
p<.10; *p<.05; **p<.01.
prospective memory performance on particular prospective memory tasksthat is not mediated by frontal/executive processes. Inthis case, age-related differences in prospective memorywill still exist after the variance in performance due tofrontal/executive processes is taken into account. Weconducted stepwise regression analyses with prospective memory performances as the dependent variables.We entered the nonexecutive measures of education:retrospective memory and health in the first step, thefrontal/executive functioning measures in the secondstep, and chronological age in the third step. This allowed us to examine whether executive functioning explains variance in prospective memory performanceabove and beyond the influence of nonexecutive measures. In addition, our procedure allows us to examine ifthe increase in explained variance (A R2) due to addingage to the equation is significant. The results are displayed in Table 4. The results indicate that, with the exception of the RBMT, individual differences in executive functioning explain a significant amount of variance of prospective memory performance even after controlling for the influence of nonexecutive measures. In contrast, adding age to the regression does not lead to a significant increase in explained variance in performance in the laboratory event- and time-based prospective memory tasks. However, regarding the complex prospective memory task all three predictor groups—nonexecutive measures, executive measures, and age—did significantly contribute to the explanation of 77% of the variance. The results are also supported by four separate ANOVA analyses with the three groups of young, old/high executive functioning, and old/low executive functioning as independent variables and the four prospective measures as the dependent variables. There was no effect of group membership for the RBMT task, F(2, 75)=1.35, p>.25, or for the eventbased task, the time-based task, and the complex memory task, F(2, 75)=4.75, p<.05; F(2, 75)=4.30, p<.05; and F(2, 75)= 69.98, p>.001, respectively. Posttests reveal no significant mean differences in the RBMT task, similar performances in the young and old/high executive functioning vs. old/low executive functioning groups in the event- and time-based task, and significant differences between all three groups in the complex prospective memory task. DISCUSSION A recent focus in prospective memory research has been to consider the role of frontally mediated executive functions for prospective remembering (Burgess, 2000b; Burgess et al., 2000; Glisky, 1996; McDaniel et al., 1999). However, the findings are disparate, with some studies reporting relations between frontal processes and prospective remembering (e.g., Burgess, 2000a; McDaniel et al.,
PROSPECTIVE MEMORY PERFORMANCE
17
1999), whereas others do not necessarily implicate frontal processes (Cockburn et al., 2000; Mi. Martin et al., TABLE 4 Regression summaries for four prospective memory performance measures as dependent variables Prospective memory measures RBMT Predictors R ∆R2 Step 1 Nonexecutive measures .01 .01 Step 2 Executive functioning .04 .03 score Step 3 Age .10 .06 *p <.05; ** <.01; ***p <.001. 2
Event-Based Time-Based MTPM/Performance R
2
∆R2
R
2
∆R2
R
2
∆R2
.05 .05
.06 .06
.51*** .51***
.27** .22**
.22* .17*
.62*** .11*
.29** .02
.22* .00
.77*** .15***
2000). Therefore, the first goal of our study was to examine and clarify the contribution of prefrontal executive processes to performance in four standard prospective memory tasks. These tasks were taken from the experimental and clinical literature, and differ with respect to the hypothesized involvement of executive processes of planning, inhibition, and self-initiated monitoring. In particular, whereas the RBMT emphasizes the retention and reinstantiation phases, the standard time-and event-based tasks focus on the intention execution phase. Finally, the complex prospective paradigm focuses on both the intention formation and the intention execution phases, thus theoretically making the largest demands on executive processes. On the part of the executive processes, mixed results concerning the involvement of executive processes in prospective memory performance reported in the literature could have been caused by the use of a particular measure focusing on a particular executive function. Therefore, in our study, we used a broadly defined construct of executive functioning by including measures of planning, inhibition, monitoring, and cognitive and response flexibility. The results improve our understanding of the neuropsychological processes involved in prospective memory performance in two important ways. First, we had expected increasing age differences in prospective memory performance the more executive functions are hypothesized to be involved, i.e., the more the intention formation and intention execution phases of the prospective task are emphasized. Based on the idea that frontal functioning may show preferential decline with age (e.g., Schretlen et al., 2000; West, 1996), we suggested that agerelated decline in prospective memory will be observable in prospective tasks that are shown to be significantly associated with frontal/ executive functioning.
18 MARTIN, KLIEGEL, MCDANIEL
Therefore, we expected age-related deficits in frontal/executive functions and in prospective memory performance across the indicators for each construct. The results clearly demonstrate age-related differences in prospective memory performance in three of the four indicators of prospective memory performance. Only the Remembering-a-Belonging subtest from the Rivermead Behavioral Memory Test, a prospective memory test taken from the assessment literature, did not reveal an age-related performance difference. Similarly, there are agerelated differences in three indicators of frontal/ executive functioning. One may conclude that there is a clear relationship between the amount of involvement of frontal/executive functions and the degree to which age effects occur. This was supported by the results from ANOVAs examining differences between the groups of young, old/ high executive functioning, and old/low executive functioning. Our finding may help to explain the disparate findings from the prospective memory literature: Studies finding no age effects might have used prospective memory tasks focusing mainly on the retention and/or reinstantiation phase of the tasks, thus requiring only a minimal involvement of frontal/executive functions (such as the Remembering-a-Belonging test), whereas studies finding age differences might have used prospective tasks with a larger involvement of frontal/executive functions. For example, in studies where participants were asked to send back postcards on specified dates and were allowed to use the postcards as external reminders (e.g., Patton & Meit, 1993; West, 1988), or in a number of event-based tasks not requiring strategic monitoring of the task environment (Einstein & McDaniel, 1990), there were no age differences in performance (cf. McDaniel & Einstein, 2000, for a theoretical discussion on the role of strategic monitoring in prospective memory performance). In the light of the present results, this is not surprising, assuming that external cues in a familiar environment minimize the need to plan to execute an intention (one relies on the efficiency of the cue) or to monitor a complex environment while trying to realize the delayed intention, i.e., to engage in frontal/executive processing. In contrast, studies in which participants had to switch between a number of different prospective actions or were engaged in demanding ongoing task activities (e.g., Kliegel et al., 2000; Mi. Martin & Schumann-Hengsteler, 2001; Maylor, 1996; Park et al., 1997) showed age differences. Given a large number and high complexity of task demands, this could be explained by assuming that the execution of the delayed intentions in these studies requires the use of frontal/executive functions like planning, attention switching, or inhibition. Second, although the near-ceiling performance of the younger adults prevents us from making strong statements about potential age differences, when the correlations between executive functioning and prospective memory performance are compared between young and old adults, it seems that interindividual differences in executive functioning are particularly predictive of prospective memory performance in old age, but not in young adults. This might be because executive processes become increasingly important as the overall
PROSPECTIVE MEMORY PERFORMANCE
19
level of cognitive processes declines in old age (e.g., Verhaeghen & Salthouse, 1997). Age differences in the correlation between executive functioning and prospective memory would, thus, become obvious in prospective memory tasks making relatively few demands for storage and cued retrieval of material, but, instead, requiring planning, inhibition, and attention switching. Therefore, a high level of prospective memory performance in older adults compared to young adults seems to depend more strongly on effective executive processes. Another explanation could be that different constellations of executive processes are involved in prospective memory performance in young and old adults. This has been suggested by Bisiacchi (1996) and has recently been supported by fMRI studies showing that different activations are present for younger and older adults on identical working memory tasks (Mitchell, Johnson, Raye, & D’Esposito, 2000; Rypma & D’Esposito, 2001). However, results are precursory, and only future research combining more difficult prospective tasks to prevent ceiling performance in the younger adults and a broader spectrum of executive functioning measures might be able to examine the executive processes involved in prospective memory performance in more detail. The second goal of our study was to examine if individual differences in frontal/executive functions rather than age per se explain most of the age-related variance in prospective memory performance (e.g., Cherry & LeCompte, 1999; Kliegel et al., 2000). In other words, we examined if, after controlling for the effects of nonexecutive influences such as education, health, or retrospective memory, significant portions of the age-related differences in prospective memory performance are eliminated once the variance in performance due to frontal/executive process is taken into account, or if, alternatively, some age deficits remain even after controlling for nonexecutive skills and after factoring out effects of frontal/executive processes. The results demonstrate that a substantial proportion of the variance in prospective memory performance can be explained by differences in executive functioning above and beyond the influence of nonexecutive functions. This was at least true for the event-based task, the time-based task, and the MTPM task. There was no significant predictor for the performance in the simple single-task RBMT measure. In addition, regarding the most complex multitask prospective memory measure, adding age to the regression equation further improved the prediction of prospective memory performance. In this paradigm only, nonexecutive measures did explain a significant amount of the prospective performance at first. This suggests that frontal/executive functioning is an important predictor for prospective memory performance, even after controlling for age differences in health, education, and retrospective memory. Moreover, age did not contribute significantly to the prediction of prospective memory in two experimental paradigms testing time-based and event-based prospective memory performance after controlling for executive functioning. Thus, for these tasks, there is no support for the assumption that prospective memory signals age-related decline that is differentiated from typical markers of age-related
20 MARTIN, KLIEGEL, MCDANIEL
frontal decline (Wecker et al., 2000). In contrast, results from the most complex experimental paradigm suggest that individual differences in nonexecutive measures, individual differences in frontal/executive functioning, and individual age differences all contribute to the prediction of prospective memory performance. Hence, in this paradigm, there is support for the assumption that prospective memory signals age-related decline that is differentiated from typical markers of age-related frontal decline. As an explanation, one might offer that in the complex prospective memory task overall processing demands are most diverse and at a relatively high level. Thus, a number of different nonexecutive processes might be involved, and their predictive power in explaining performance becomes substantial. For example, this may be true for acquired problem-solving strategies that might support performance in the complex prospective memory task. Still, to perform well in this prospective memory task also requires the on-line processes of planning, inhibition, and switching between the tasks. Thus, executive processes play an important role in predicting prospective memory performance even after controlling for individual differences in nonexecutive measures. Our results indicate that even after controlling for age-related declines in nonexecutive and executive processes, there are still age-related differences in processes involved in complex prospective memory performance that have not been captured by our instruments. There are at least two possible explanations of this finding. First, there might be other nonexecutive factors, e.g., motivational or emotional processes, that are contributing to age-related differences in complex prospective memory performance that have not been examined in our study (cf. Kliegel et al., 2001, 2002b, this issue). Second, there might be executive processes other than the ones examined in our three standard measures that are important for complex prospective memory performance. As a consequence, our findings suggest a need for measurement instruments capturing the full range of nonexecutive and executive processes potentially related to complex prospective memory performance. Based on our data, reasoning about the exact mechanisms responsible for the (age-related) interplay of nonexecutive processes, frontal/executive functioning, and prospective memory performance has to remain speculative at this point. Further research needs to disentangle the nonexecutive and frontal/ executive processes that might be differentially related to different types of prospective tasks. Overall, it seems clear that frontal/executive functions are related to prospective memory performance across a range of prospective paradigms. It also seems clear that age differences in prospective memory performance partially depend on age-related individual differences in frontal/executive functions. Further research involving complex prospective memory tasks is now needed to examine the exact mechanisms responsible for these effects (Ellis & Kvavilashvili, 2000).
PROSPECTIVE MEMORY PERFORMANCE
21
REFERENCES Bisiacchi, P.S. (1996). The neuropsychological approach in the study of prospective memory. In M.A.Brandimonte, G.O.Einstein, & M.A.McDaniel (Eds.), Prospective memory: Theory and applications ( pp. 297–317 ). Mahwah, NJ: Lawrence Erlbaum Associates Inc. Bisiacchi, P.S. (July 2000). Prospective memory in Parkinson’s disease . Paper presented at the 1st International Conference on Prospective Memory, Hatfield, UK. Burgess, P.W. (July 1996). Prospective memory following frontal lobe damage . Paper presented at the 2nd International Conference on Memory, Abano-Padua, Italy. Burgess, P.W. (July 2000a). The cognitive neuroscience of prospective memory . Paper presented at the 1st International Conference on Prospective Memory, Hatfield, UK. Burgess, P.W. (2000b). Strategy application disorder: The role of the frontal lobes in human multitasking. Psychological Research , 63 , 279–288 . Burgess, P.W., Veitch, E., de Lacy Costello, A., & Shallice, T. (2000). The cognitive and neuroanatomical correlates of multi-tasking. Neuropsychologia , 38 , 848–863 . Cherry, K.E., & LeCompte, D.C. (1999). Age and individual differences influence prospective memory. Psychology and Aging , 14 , 60–76 . Cockburn, J., Keene, J., & Hope, T. (July 2000). Are prospective memory measures of the RBMT more vulnerable to cognitive impairment than retrospective memory measures ? Paper presented at the 1st International Conference on Prospective Memory, Hatfield, UK. Cockburn, J., & Smith, P.T. (1988). Effects of age and intelligence on everyday memory tasks. In M.M. Gruneberg, P.E.Morris, & R.N.Sykes (Eds.), Practical aspects of memory: Current research and issues. Vol. 2 : Clinical and educational implications ( pp.132–136 ). Chichester, UK: Wiley. Cockburn, J., & Smith, P.T. (1991). The relative influence of intelligence and age on everyday memory. Journal of Gerontology , 46 , P31–P36 . Cockburn, J., & Smith, P.T. (1994). Anxiety and errors of prospective memory among elderly people. British Journal of Psychology , 85 , 273–282 . Dempster, F.N. (1992). The rise and fall of the inhibitory mechanism: Toward a unified theory of cognitive development and aging. Developmental Review , 12 , 45–75 . Einstein, G.O., & McDaniel, M.A. (1990). Normal aging and prospective memory. Journal of Experimental Psychology: Learning, Memory, and Cognition , 16 , 717– 726 . Einstein, G.O., & McDaniel, M.A. (1996). Retrieval processes in prospective memory: Theoretical approaches and some new empirical findings. In M.A.Brandimonte, G.O.Einstein, & M.A.McDaniel (Eds.), Prospective memory: Theory and applications ( pp. 115–141 ). Mahwah, NJ: Lawrence Erlbaum Associates Inc. Einstein, G.O., McDaniel, M.A., Richardson, S.L., Guynn, M.J., & Cunfer, A.R. (1995). Aging and prospective memory: Examining the influences of self-initiated retrieval processes. Journal of Experimental Psychology: Learning, Memory, and Cognition , 21 , 996–1007 . Einstein, G.O., Smith, R.E., McDaniel, M.A., & Shaw, P. (1997). Aging and prospective memory: The influence of increased task demands at encoding and retrieval. Psychology and Aging , 12 , 479–488 .
22 MARTIN, KLIEGEL, MCDANIEL
Ellis, J. (1996). Prospective memory or the realization of delayed intentions: A conceptual framework for research. In M.A.Brandimonte, G.O.Einstein, & M.A.McDaniel (Eds.), Prospective memory: Theory and applications ( pp. 1–22 ). Mahwah, NJ: Lawrence Erlbaum Associates Inc. Ellis, J. & Kvavilashvili, L. (2000). Prospective memory in 2000: Past, present, and future directions. Applied Cognitive Psychology , 14 , S1–S9 . Evans, J., & Wilson, B.A. (1992). A memory group for individuals with brain injury. Clinical Rehabilitation , 6 , 75–81 . Glisky, E.L. (1996). Prospective memory and the frontal lobes. In M.A.Brandimonte, G.O.Einstein, & M.A.McDaniel (Eds.), Prospective memory: Theory and applications ( pp. 249–266 ). Mahwah, NJ: Lawrence Erlbaum Associates Inc. Greve, K.W., Brooks, J., Crouch, J.A., & Williams, M.C. (1997). Factorial structure of the Wisconsin Card Sorting Test . British Journal of Clinical Psychology , 36 , 283–285 . Guynn, M.J., McDaniel. M.A., & Einstein, G.O. (2001). Remembering to perform actions: A different type of memory? In H.D.Zimmer, R.L.Cohen, M.J.Guynn, J.Engelkamp, R.Kormi-Nouri, & M.A.Foley (Eds.), Memory for action: A distinct form of episodic memory? ( pp. 25–48 ). New York: Oxford University Press. Heaton, R.K., Chelune, G.J., Talley, J.L., Kay, G.G., & Curtiss, G. (1993). Wisconsin Card Sorting Test Manual . Odessa, FL: Psychological Assessment Resources. Hon, J., Huppert, F.A., Holland, A.J., & Watson, P. (1998). The value of the Rivermead Behavioral Memory Test (Children’s Version) in an epidemiological study of older adults with Down syndrome. British Journal of Clinical Psychology , 37 , 15–29 . Houx, P.J., Jolles, J., & Vreeling, F.W. (1993). Stroop interference: Aging effects assessed with the Stroop colorword test . Experimental Aging Research , 19 , 209– 224 . Huppert, F.A. & Beardsall, L. (1993). Prospective memory impairment as an early indicator of dementia. Journal of Clinical and Experimental Neuropsychology , 15 , 805–821 . Huppert, F.A., Johnson, T., & Nickson, J. (2000). High prevalence of prospective memory impairment in the elderly and in early-stage dementia: Findings from a populationbased study . Applied Cognitive Psychology , 14 , S63–S81 . Kliegel, M., Martin, M., McDaniel, M.A., & Einstein, G. O. (2001). Varying the importance of a prospective memory task: Differential effects across time- and eventbased prospective memory. Memory , 9 , 1–11 . Kliegel, M., Martin, M., McDaniel, M.A., & Einstein, G. O. (2002a). Complex prospective memory and executive control of working memory: A process model. Psychologische Beiträge , 44 , 303–318 . Kliegel, M., Martin, M., McDaniel, M.A., & Einstein, G. O. (April 2002b). Prospective memory and aging: How planning affects performance . Paper presented at the Cognitive Aging Conference in Atlanta, GA, USA. Kliegel, M., Martin, M., McDaniel, M.A., & Einstein, G. O. (in press). Why do we remember to perform intended actions? Importance effects on performance in eventbased prospective memory. Memory . Kliegel, M., McDaniel, M.A., & Einstein, G.O. (2000). Plan formation, retention, and execution in prospective memory: A new approach and age-related effects. Memory and Cognition , 28 , 1041–1049 .
PROSPECTIVE MEMORY PERFORMANCE
23
Kopp, U., & Thöne, A.I.T. (July 2000). Role of executive functions and memory processes in delayed intentions after head injury . Paper presented at the 1 st International Conference on Prospective Memory, Hatfield, UK. Kvavilashvili, L. (1987). Remembering intention as a distinct form of memory. British Journal of Psychology, 78 , 507–518 . Kvavilashvili, L., & Ellis, J. (1996). Varieties of intention: Some distinctions and classifications. In M.Brandimonte, G.O.Einstein, & M.A.McDaniel (Eds.), Prospective memory: Theory and applications ( pp. 23–51 ). Mahwah, NJ: Lawrence Erlbaum Associates Inc. Martin, Ma. (1986). Ageing and patterns of change in everyday memory and cognition. Human Learning , 5 , 63–74 . Martin, Mi., Kliegel, M., McDaniel, M.A., & Einstein, G. O. (July 2000). The neuropsychology of prospective memory: The role of central executive functions . Paper presented at the 1st International Conference on Prospective Memory, Hatfield, UK. Martin, Mi., & Schumann-Hengsteler, R. (2001). How task demands influence time-based prospective memory performance in young and older adults. International Journal of Behavioural Development , 25 , 386–391 . Maylor, E.A. (1993). Aging and forgetting in prospective and retrospective memory tasks. Psychology and Aging , 8 , 420–428 . Maylor, E.A. (1996). Does prospective memory decline with age? In M.A.Brandimonte, G.O.Einstein, & M.A.McDaniel (Eds.), Prospective memory: Theory and applications ( pp. 173–198 ). Mahwah, NJ: Lawrence Erlbaum Associates Inc. McDaniel, M.A., & Einstein, G.O. (1993). The importance of cue familiarity and cue distinctiveness in prospective memory. Memory , 1 , 23–41 . McDaniel, M.A. & Einstein, G.O. (2000). Strategic and automatic processes in prospective memory retrieval: A multiprocess framework. Applied ed Cognitive Psychology , 14 , S127–S144 . McDaniel, M.A., Glisky, E.L., Rubin, S.R., Guynn, M.J., & Routhieaux, B.C. (1999). Prospective memory: A neuropsychological study. Neuropsychology , 13 , 103–110 . McDaniel, M.A., Robinson-Riegler, B., & Einstein, G.O. (1998). Prospective remembering: Perceptually-driven or conceptuallydriven processes? Memory and Cognition , 26 , 121–134 . Mitchell, K.J., Johnson, M.K., Raye, C.L., & D’Esposito, M. (2000). FMRI evidence of age-related hippocampal dysfunction in feature binding in working memory. Cognitive Brain Research , 10 , 197–206 . Morris, R.G., Evenden, J.L., Sahakian, B.J., & Robbins, T. W. (1987). Computer-aided assessment of dementia: Comparative studies of neuropsychological deficits in Alzheimer-type dementia and Parkinson’s disease. In S. M.Stahl & S.D.Iversen (Eds.), Cognitive neurochemistry ( pp. 21–36 ). New York: Oxford University Press. Park, D.C, Hertzog, C, Kidder, D.P, Morrell, R.W., & Mayhorn, C. (1997). The effect of age on event-based and time-based prospective memory. Psychology and Aging , 12 , 314–327 . Park, D.C, & Kidder, D.P. (1996). Prospective memory and medication adherence. In M.A.Brandimonte, G.O. Einstein, & M.A.McDaniel (Eds.), Prospective memory: Theory and applications ( pp. 369–390 ). Mahwah, NJ: Lawrence Erlbaum Associates Inc.
24 MARTIN, KLIEGEL, MCDANIEL
Patton, G.W.R. & Meit, M. (1993). Effect of aging on prospective and incidental memory. Experimental Aging Research , 19 , 165–176 . Roediger, H.L. III (1996). Prospective memory and episodic memory. In M.A.Brandimonte, G.O.Einstein, & M.A. McDaniel (Eds.), Prospective memory: Theory and applications ( pp. 149–155 ). Mahwah, NJ: Lawrence Erlbaum Associates Inc. Rypma, B., & D’Esposito, M. (2001). Age-related changes in brain-behaviour relationships: Evidence from event-related functional fMRI studies. European Journal of Cognitive Psychology , 13 , 235–256 . Schretlen, D., Pearlson, G.D., Anthony, J.C, Aylward, E. H., Augustine, A.M., Davis, A., & Barta, P. (2000). Elucidating the contributions of processing speed, executive ability, and frontal lobe volume to normal age-related differences in fluid intelligence. Journal of the International Neuropsychological Society , 6 , 52–61 . Shallice, T., & Burgess, P.W. (1991). Deficits in strategy application following frontal lobe damage in man. Brain , 114 , 727–741 . Shimamura, A.P, Janowsky, J.S., & Squire, L.R. (1991). What is the role of frontal lobe damage in memory disorders? In H.S.Levin, H.M.Eisenberg, & A.L. Benton (Eds.), Frontal lobe function and dysfunction ( pp. 173–195 ). New York: Oxford University Press. Spieler, D.H., Balota, D.A., & Faust, M.E. (1996). Stroop performance in healthy younger and older adults and in individuals with dementia of the Alzheimer’s type. Journal of Experimental Psychology: Human Perception and Performance , 22 , 461–419 . Stroop, J.R. (1935). Studies of interference in serial verbal reactions . Journal of Experimental Psychology , 18 , 643–662 . Stuss, D.T., & Benson, D.F. (1987). The frontal lobes and control of cognition and memory. In E.Perecman (Ed.), The frontal lobes revisited ( pp. 141–158 ). Hillsdale, NJ: Lawrence Erlbaum Associates Inc. Verhaeghen, R, & Salthouse, T.A. (1997). Meta-analyses of age-cognition relations in adulthood: Estimates of linear and nonlinear age effects and structural models. Psychological Bulletin , 122 , 231–249 . Ward, G., & Allport, A. (1997). Planning and problemsolving using five-disc Tower of London task. Quarterly Journal of Experimental Psychology: Human Experimental Psychology , 50A , 49–78 . Wecker, N.S., Kramer, J.H., Wisniewski, A., Delis, D.C, & Kaplan, E. (2000). Age effects on executive ability. Neuropsychology , 14 , 409–414 . West, R.L. (1988). Prospective memory and ageing. In M.M. Gruneberg, P.E.Morris, & R.N.Sykes (Eds.), Practical aspects of memory: Current research and issues , Vol. 2 : Clinical and Educational Implications ( pp. 119–125 ). Chichester, UK: Wiley. West, R.L. (1996). An application of prefrontal cortex function theory to cognitive aging. Psychological Bulletin , 120 , 272–292 . Wilson, B.A., Cockburn, J., & Baddeley, A.D. (1985). The Rivermead Behavioral Memory Test . Titchfield, UK: Thames Valley Test Co. Wiseman, K.A., Ratcliff, G., Chase, S., Laporte, D.J., Robertson, D.U., & Colantonio, A. (2000). Does a test of functional memory during the post-acute period predict longterm outcome of traumatic brain injury? Brain and Cognition , 44 , 14–18 .
Prospective memory and ageing : Is task importance relevant? Matthias Kliegel, Mike Martin, and Caroline Moor University of Zurich, Switzerland
Memory for activities to be performed in the future, i.e., prospective memory, such as remembering to take medication or remembering to give a colleague a message, is a pervasive real-world memory task that has recently begun to attract the attention of numerous researchers. Age effects in prospective memory have been found particularly in complex paradigms requiring participants to remember to switch between several subtasks in a limited time period (e.g., Kliegel, McDaniel, & Einstein, 2000). Here, most of the older adults try to complete one or two subtasks and to forget the prospective instruction to work on all subtasks. Since recent findings in this context show that one profits from tips regarding the relevant task in complex double tasks, it seems likely that age effects in prospective memory tasks might also be due to the lack of information about the importance of the prospective task. To test this hypothesis, the importance of the prospective task was varied in the present study with 104 young and old participants by providing motivational incentives to interrupt and switch during the introduction phase (plan formation) as well as during the execution phase. Also, interindividual differences regarding nonexecutive as well as executive cognitive resources were analysed, thus allowing estimation of the relationship between these factors and (age-related) performance in complex prospective remembering. The results show age effects in favour of the younger group in all task components of the complex prospective multitask. In contrast, there was no effect of the present experimental manipulation of motivational incentives. Finally, in regression analyses, intention formation, in particular, was found to be a significant predictor of intention execution, explaining most of the age-related variance. In sum, our results specifically highlight the fundamental importance of adequately planning the complex intention and do not support the hypothesis that age-related decrements in performance are reflecting a lack of motivation in the present complex prospective memory paradigm.
26 KLIEGEL, MARTIN, MOOR
La mémoire pour les activités qui seront exécutées dans le futur, c’est-à-dire la mémoire prospective, comme se rappeler de prendre des médicaments ou se rappeler de transmettre un message à un collègue, est une tâche mnémonique envahissante concrete qui a récemment commence a attirer l’attention de nombreux chercheurs. Les effets de l’âge sur la mémoire prospective ont été particulièrement observés dans les paradigmes complexes requérant, pour les participants, de se rappeler de transférer entre plusieurs soustâches dans une période de temps limitée (p. ex., Kliegel, McDaniel et Einstein, 2000). La plupart des adultes plus ages ont tendance a vouloir compléter une ou deux sous-tâches et à oublier la directive prospective pour travailler sur toutes les sous-tâches. Dans la perspective que les résultats récents dans ce contexte montrent qu’un individu profite des conseils relatifs à saillance de la tâche dans les tâches doubles complexes, il semble probable que les effets de l’âge dans les tâches de mémoire prospective peuvent également être dus au manque d’information se rapportant a la saillance de la tâche prospective. Pour vérifier cette hypothèse, cette étude fut menée auprès de 104 participants jeunes et ages auprès desquels on faisait varier la saillance de la tâche prospective en fournissant des incitatifs motivationnels pour interrompre et transférer durant la phase d’introduction (formation du plan) tout comme durant la phase d’exécution. De plus, les differences interindividuelles concernant les ressources cognitives executives et non executives furent analysées, permettant d’estimer la relation entre ces facteurs (relativement a l’âge) et la performance dans le rappel prospectif complexe. Les résultats montrent que les effets de l’âge sont favorables pour le groupe de jeunes participants, dans toutes les composantes de la multi-tâche prospective complexe. Toutefois, la
© 2003 International Union of Psychological Science http:// DOI: 10.1080/ www.tandf.co.uk/ 00207590244000205 journals/pp/ 00207594.html
Requests for reprints should be addressed to Matthias Kliegel, Institute of Psychology, Department of Gerontopsychology, University of Zurich, Schaffhauserstr. 15, CH-8006 Zürich, Switzerland (E-mail:
[email protected]). Part of the preparation of this manuscript was supported by DFG grant Ma 1896/4–1.
PROSPECTIVE MEMORY AND AGEING
manipulation expérimentale des incitatifs motivationnels n’a pas eu d’effet significatif pour aucun des deux groupes. Finalement, dans les analyses de regression, la planification (c.-à-d. la formation de l’intention) fut particulièrement soulevée comme prédicteur significatif de l’intention d’exécution, expliquant la plus grande part de la variance reliée a l’âge. En somme, les résultats mettent en lumière l’importance fondamentale de planifier adéquatement l’intention complexe et de ne pas supporter l’hypothèse selon laquelle les reductions de la performance reliées a l’âge reflètent un manque de saillance de la tâche dans le paradigme complexe de la mémoire prospective. El recuerdo de actividades a realizarse en un futuro, es decir, la memoria prospectiva, tal como recordar el tomar un medicamento o el dar a un colega un mensaje, es una tarea de memoria dominante real que recientemente ha atraído la atención de numerosos investigadores. Se ha encontrado efectos de la edad en la memoria prospectiva en paradigmas complejos que requieren que los participantes recuerden cambiar entre varias subtareas en un periodo de tiempo limitado (por ejemplo, Kliegel, McDaniel, & Einstein, 2000). En ese estudio, la mayoría de los adultos mayores tendían a tratar de terminar una o dos subtareas y a olvidar la instrucción prospectiva de trabajar en todas las subtareas. Dado que los hallazgos recientes en este contexto muestran que es posible aprovechar varios consejos respecto a la prominencia de las tareas relevantes en las tareas dobles complejas, sería probable que los efectos de la edad en las tareas de memoria prospectiva se debieran también a la falta de información acerca de la prominencia de la tarea prospectiva. Para someter a prueba esta hipótesis, en el presente estudio se varió la prominencia de la tarea prospectiva a 104 participantes jóvenes y ancianos al proporcionarles incentivos motivacionales para que interrumpieran y cambiaran durante la fase introductoria (formación del plan) así como durante la fase de ejecución. También, se analizó las diferencias entre los individuos respecto a sus recursos cognitivos tanto ejecutivos como no ejecutivos, lo que permitió calcular la relación entre estos factores (relativos a la edad) y el desempeño en el recuerdo prospectivo complejo. Los resultados mostraron efectos de la edad que favorecen al grupo joven en todos los componentes de la multitarea prospectiva compleja. En contraste, ninguno de los grupos sacó provecho significativamente de la manipulación experimental de incentivos motivacionales. Finalmente, se encontró del análisis de regresión que la planeación (es decir, la formación de la intención) predecía de manera significativa la intención de ejecución, al explicar la mayor parte de la varianza relativa a la edad. En resumen, los resultados
27
28 KLIEGEL, MARTIN, MOOR
resaltan específicamente la importancia de la planeación adecuada de la intención compleja y no apoyan la hipótesis de que la reducción en el desempeño relativa a la edad refleja una falta de prominencia de la tarea en el paradigma complejo de memoria prospectiva. In everyday life, remembering to perform actions planned earlier can be highly relevant, especially when one is busy with a competing activity at the same time (Brandimonte, Einstein, & McDaniel, 1996). This is one of the reasons why the number of studies on prospective memory has rapidly increased over the last 10 years (cf. Ellis & Kvavilashvili, 2000). Even though prospective memory, in order to master daily routine, is required at virtually every stage during a lifespan (cf. Martin & Kliegel, 2003), in old age, in particular, an intact prospective memory has been found to be a major prerequisite for successful independent living (Kliegel, in press; Martin, 2001). Therefore, researchers’ recent interest focuses on factors able to explain age differences in prospective remembering (Einstein, McDaniel, Manzi, Cochran, & Baker, 2001; Ellis & Kvavilashvili, 2000; Kliegel, McDaniel, & Einstein, 2000; Martin, Kliegel, & McDaniel, this issue). Current theories of age effects in prospective remembering refer to the structure of the prospective task as well as the involved cognitive processes. For example, Kvavilashvili and Ellis (1996) postulate a range of prospective task characteristics believed to be related to differences in performance. A rather critical factor seems to be the importance of the prospective task, which has been given close empirical attention over recent years. Several studies with younger adults and simple prospective memory tasks have demonstrated that increasing the prospective task’s importance will lead to better prospective memory performance (Kliegel, Martin, McDaniel, & Einstein, 2001; in press; Kliegel & Martin, in press; Kvavilashvili, 1987; McDaniel & Einstein, 2000). Since recent findings in this context show that older adults, in particular, profit from tips regarding the relevant task in complex double tasks (e.g., Hohaus, Shum, & White, 2000), it seems likely that age effects might also be due to the lack of information about the importance of the prospective task. In other words, the importance of actually completing all of the prospective actions planned beforehand at the proper time may be less obvious to older adults than to younger ones. So far, though, there are no empirical proofs for this hypothesis. Apart from the idea just discussed, it has further been proposed that several cognitive resources may influence age-related prospective memory performance. While the influence of nonexecutive cognitive resources such as, for example, verbal intelligence (cf. Cherry & LeCompte, 1999) or retrospective memory (cf. Einstein, Holland, McDaniel, & Guynn, 1992) has been investigated repeatedly over the last years, more recent approaches discuss the effects that executive control processes like planning or inhibition (cf. Baddeley, 1996; Kliegel, Martin, McDaniel, & Einstein, 2002; Martin et al., this issue; Martin & SchumannHengsteler, 2001; Smith & Jonides, 1999) might have, particularly with respect
PROSPECTIVE MEMORY AND AGEING
29
to complex (i.e., multiple) prospective memory tasks. Unlike simple prospective laboratory tasks, where participants are usually just expected to keep in mind and later realize the intention to perform one single action, several current studies investigate so-called multiple or “complex” (Kliegel et al., 2000; Ellis & Kvavilashvili, 2000) prospective memory tasks. This paradigm requires participants to completely and on their own perform a delayed intention to switch between several similar (sub)tasks at a certain point once the intention is reinstantiated (Burgess, Veitch, de Lacy Costello, & Shallice, 2000; Kliegel et al., 2000,2002; Kvavilashvili, 1992; Kvavilashvili & Ellis, 1996; Rendell & Craik, 2000). As a complement to many traditional and fairly simple research paradigms that mainly examine the phases concerning the formation of a prospective intention or its reinstantiation (cf. Brandimonte et al., 1996), quantitatively complex prospective multitasks focus—after the phases of intention formation, retention and reinstantiation—on the actual performance of the multiple intention. Particularly in this last phase, strong age effects that are largely independent from reinstantiation (i.e., the point where one remembers and begins with the prospective task) have repeatedly been found (Kliegel et al., 2000,2002; see also Rendell & Craik, 2000). While most simple prospective paradigms do not seem to involve active planning processes (Bisiacchi, 1996), the delayed execution of an intention to switch between several tasks is believed to be related to active planning during the intention formation phase. In addition, age effects in complex prospective memory tasks are assumed to correlate with older adults’ deficits in planning these tasks (Kliegel et al., 2000,2002). A further cognitive resource to be mentioned in connection with the described age effects is inhibitory efficiency. In addition to the general finding of older people’s less effective inhibitory mechanisms when performing a cognitive task (Hasher & Zacks, 1988), it seems that, apart from planning complex prospective multiple intentions, older adults also have more difficulties in implementing encoded plans adequately and in keeping those plans in their working memory while performing competing activities. Therefore, first theoretical speculations suspect that age differences in inhibitory functioning during intention execution would be responsible for this effect (Martin & Schumann-Hengsteler, 2001; McDaniel, Glisky, Rubin, Guynn, & Routhieaux, 1999). Thus, the goal of the present study was to examine (1) importance as an indirect motivational measure and (2) cognitive resources/planning functions, investigating their influence on or relation with age-related performance in a complex prospective multiple-task paradigm, respectively. To test our assumptions, we administered the paradigm developed by Kliegel et al. (2000), since it had consistently revealed reliable age effects for various task components in previous studies (Eschen, Kliegel, & Thöne-Otto, 2002; Kliegel et al., 2000,2002; Martin et al., this issue; Philipps, MacLeod, & Kliegel, in press). Basically, the idea of this delayed six-element task (after Shallice & Burgess, 1991) is that, following certain rules, participants have to realize the delayed execution of six intentions. Therefore, they first generate a plan of how
30 KLIEGEL, MARTIN, MOOR
to complete a set of tasks consisting of six different subtasks. Next, they are required to maintain their multiple intentions during a delay period and then, finally, to reinstantiate and perform the six subtasks on their own. The latter characteristic fundamentally distinguishes the present task from stimulustriggered multiple tasks (cf. Kliegel et al., 2000). The main difficulty with planning and executing the delayed realization of these six subtasks is that, within a certain period of time, one must work on every one of the six tasks, although not all tasks can be completely solved as each one—taken alone— would require too much time. Therefore, participants are forced to interrupt each task at a self-chosen point, so as not to forget to switch between subtasks. So far, this necessity has been emphasized by instructing participants that every one of the six tasks must be worked on at least for a short time although not all tasks could be completed, and that within each subtask earlier items would be given more points than later ones. Previous studies revealed that in spite of these hints, the majority of the older participants accounted for this interrupt/switch necessity less often in their plans and thus carried out fewer self-initiated prospective switches in their actions. To test the hypothesis that the importance of a prospective task (in this case, the importance of the described necessity to interrupt and switch activity) influences (age-related) prospective memory, the importance of the prospective task was varied in the present study by modifying the necessity to interrupt and switch during the introduction phase (plan formation) as well as during the execution phase. Also, interindividual differences regarding nonexecutive as well as executive cognitive resources were analysed, thus allowing estimation of the relationship between these factors and (age-related) performance in complex prospective remembering. METHOD Design and participants In order to address our hypotheses, we chose a 2×2×2× 2 factorial design, varying the between-subjects factors task importance during the introduction phase (high vs. normal), task importance during the execution phase (high vs. normal), with age (young vs. old) as the third factor. Thirteen participants were randomly assigned to each of the eight experimental groups. One younger and four older participants were subsequently discarded because they were not native speakers of German. The reported analyses included 51 younger participants with an average age of 25 years (SD=4.5; range 20–41) and 48 older participants with an average age of 69.7 years (SD=5.9; range 59–82). Women comprised 74. 5% of the younger and 66.7% of the older participants. The younger participants were mostly students and the older participants were recruited from an older people’s institution for continuing education. Participation was voluntary. In order to ensure comparability of groups with respect to their verbal intelligence,
PROSPECTIVE MEMORY AND AGEING
31
participants were given a vocabulary test (MWT; Lehrl, 1977). Comparing intelligence quotients of both age groups, in test version A groups did not differ, whereas in version B the older group, in fact, did even better than the younger adults (M=125.33; SD=12.61 vs. M =117.38; SD=14.45), F(l, 95)=8.31, p<.01. On the whole, the older group of participants can be regarded as equivalent to the younger group in terms of verbal intelligence. Instruments and procedure All participants were given the complex memory task by Kliegel et al. (2000), in a paper-and-pencil version that had been slightly modified in order to allow manipulating importance. The procedure consisted of three phases: (1) an introduction phase, in which participants were instructed in the complex prospective memory task and asked to develop a plan; (2) a delay phase, during which intention had to be stored and individual difference variables of cognitive resource (e.g., retrospective memory, inhibition) were assessed; and (3) a performance phase, in which the prospective intention was to be reinstantiated and executed. The introduction phase. After a general introduction, the participants were instructed in the prospective task using sample sheets. They were asked to carry out six subtasks in a 6-minute time period, in a way that would allow them to maximize their overall scores. The six subtasks comprised three different task types (word finding, solving arithmetic problems, and picture naming), divided into two similar sets (A and B). Each subtask was designed so that it would take more than 1 minute to complete. Both sets of word-finding problems (based on a German vocabulary test, MWT; Lehrl, 1977) consisted of 37 groups of five items. In each group there was a true word (e.g., conceal) and four similarly spelled or similarly sounding pseudowords (e.g., concill, cauncil, concel, caunseal). The participant’s task was to circle the actual word. Each set of arithmetic problems (A and B) contained 10 items (e.g., 300/6×4=); both sets were equivalent in difficulty. Finally, the 20 pictures in each set were pictures of common objects or symbols (e.g., a house). Here, the participant’s task was to label the picture appropriately. The participants were told that there were no perfect answers in this subtask and that they should write down the first name or title that they thought of. After explaining the subtasks, the participants were told where the material for these subtasks was stored. Then, using a rule sheet, the rules were explained to the participants. The rules were as follows: 1. Your aim is to maximize your score. (a) Earlier word groups/problems/ pictures will be given more points than later ones in each subtask. (b) You have to remember to work on each one of the six tasks at least for a short time. (c) Each of the six subtasks will be given equal weight. (As mentioned above, the purpose of giving instructions about scoring was merely to emphasize the
32 KLIEGEL, MARTIN, MOOR
necessity of switching between subtasks—cf. Kliegel et al., 2000— the number of points scored was not analysed.) 2. You are not allowed to do two subtasks (A) and (B) of the same type one after the other. 3. You will have 6 minutes time. In order to lay further emphasis on task importance, i.e., to once more underline that participants would have to switch between all tasks on their own, rule l(b) was varied (manipulation of task importance during plan ning). Half of the participants were given a supplement to rule 1 (b), which said: “You have to remember to work on each one of the six tasks at least for a short time. For each correct switch you will be given extra points.” Afterwards, participants were asked to repeat the rules and any errors or omissions were corrected. Once the participants were able to repeat the rules completely, they were told that they should start working on the six tasks by themselves after having answered the question about their date of birth in the Participant Information Questionnaire (which was shown to the participants at that point). Finally, the participants were asked to develop a plan for the prospective memory task (intention formation). This plan was reported verbally and recorded on cassette tape. In line with Kliegel et al. (2000), plan elaborateness was analyzed in terms of a score that took into account three main features: (1) the number of rules named or implicitly included in a participant’s intention for a specific step, (2) the number of specifications a participant made regarding a particular order for performing a task by giving a reason for that step, and (3) the number of executable items in the plan. To assess the number of executable items, we noted how many executable steps the participant indicated, i.e., (a) the number of task types he/she planned to initiate (words, pictures, arithmetic problems—1 point each), (b) whether he/she specified the steps concerning the version (A or B-1 point each), (c) whether he/she specified the steps concerning the time planned to spend on each task or version (1 point each) and (d) whether he/ she specified the number of items he/she planned to complete in each step (1 point each). The intentionelaboration score was the sum of the number of features (described above) included in a participant’s plan. For example, the plan “…seeing that I must not work on a version A and B of the same task type one after the other, I shall first do the version A word task, then the A pictures, then the A arithmetic problems, then the B words, then the B pictures and at the end the B arithmetic problems. This way at least I will have worked on all tasks for a short time…” would be given 12 points (six executable items as well as six specifications concerning version A and B) plus two rules (rule l(b) and rule 2 =2 points), making a total of 14 points. The theoretical minimum of the score is 0, which would indicate that the participant did not plan at all. The maximum score is, in principle, unlimited.
PROSPECTIVE MEMORY AND AGEING
33
The delay phase. In this phase, participants were busy with distractor tasks. Among other things, the participants were given a colour-word version of the Stroop task (Houx, Jolles, & Vrelling, 1993) in order to measure inhibition (e.g., Dempster, 1992). After the Stroop task, the participants had to recall their plans for the prospective memory task. Recollection served as a measure of retrospective memory for the previously formed intentions. Retrospective memory was judged by the accuracy with which a plan was recalled, relative to when it was initially stated (in per cent). The execution phase. After various other distractor tasks, the participants were requested to fill out the Participant Information Questionnaire. As mentioned above, the main objective of this study was to investigate the complete execution of the complex prospective intention (i.e., to switch between all six subtasks), which participants had to perform on their own. Participants who neither switched to the prospective memory task when they should have nor later on during the questionnaire were prompted to do so after having finished the questionnaire (see Kliegel et al., 2000, 2002; Martin & Kliegel, in press). This measure was taken to ensure that the initial reinstantiation of the set of multiple prospective intentions was comparable between participants. In order to manipulate the importance of multiple, self-initiated switches in the prospective task during the execution phase, there were two versions of task sheets. In the version expected to emphasize task importance, the amount of points that would be given for each item was explicitly indicated, whereas in the normal version it was not. In the high importance version, following rule l(a), the amount of possible points per item decreased rapidly down the list, thus making it very uneconomical to solve more than two items per subtask. Half of the participants worked on the normal version (no explicit indication of points) for 6 minutes, the other half on the ‘high-need-for-switching’ sheets (high importance condition) with points marked behind each item. In accordance with Kliegel et al. (2000; see also Burgess et al., 2000), in tention-execution was measured by the number of started tasks (out of six possible ones). Afterwards, the participants finished the questionnaire. Then, those who had not completed all word-finding problems during the prospective memory task were asked to do so at that point (no time limits). At the end of the experiment participants were debriefed. RESULTS Age and importance We performed 2×2×2 ANOVAS to determine whether the between-subjects factors of age (young vs. old), importance during introduction phase (high vs. normal), and importance during execution phase (high vs. normal) had any
34 KLIEGEL, MARTIN, MOOR
Figure 1 . Effects of age and high vs. normal importance (during the planning and the performance phase) on intention execution.
influence on prospective memory measures such as plan complexity, plan recall, and intention execution. Plan complexity Results showed a significant age effect; younger participants developed more complex plans than older ones (M=15.19; SD=6.45 vs. M=8.17; SD=6.73), F(1, 91)=26.88, p <.01, η2=.23. However, no differences were found when task importance was manipulated during the introduction phase (high-importance: M= 10.98; SD=7.45, vs. normal-importance: M=12.58; SD =7.43), F(l, 91)=1.36, p<. 3, η2=.02, or during the execution phase (high-importance: M=11.39; SD=7.47, vs. normal-importance: M=12.16; SD=7.48), F(l, 91)= 0.16, p<.7, η2=.002. There were no significant interactions between factors (all Fs< 1). Plan recall Younger participants recalled their plans slightly, though significantly, better than did older ones (M =94.36%; SD=14.37 vs. M=79.25%; SD=35.49), F(1, 75) =6.38; p<.01, η2=.08. As with plan complexity, there were neither reliable effects of task importance nor significant interactions between factors (all Fs<1; all η2<. 01). Intention execution Overall, the participants carried out an average of 4.63 subtasks (SD=1.58). As Figure 1 indicates, younger adults were found to perform more correct switches between subtasks than older ones (M=5.47; SD=1.05 vs.M=3.73; SD=1.55), F(l, 91)=41.57, p<.01,η2=.31. Neither salience condition had any effects on the number of performed tasks. Further, no reliable interactions were found between factors (all F<1; all η2<.01).
PROSPECTIVE MEMORY AND AGEING
35
Cognitive resources The second objective of this study was to examine whether individual differences in verbal intelligence, retrospective memory, planning, or inhibition can explain parts of the (age-related) variance in executing the complex prospective memory task. For evaluation, we conducted two regression analyses on intention execution as the dependant measure (after Salthouse, 1991; see Kliegel et al., 2000; Martin et al., this issue). In the first regression equation, age as the only predictor explained 32% of the variance. In order to examine whether nonexecutive or executive measures were associated with this age-related variance, we subsequently conducted a further hierarchical regression analysis. As predictors, the nonexecutive measures verbal intelligence and retrospective memory were included in a first step. In a second step, we included the executive measures planning and inhibition and, in step three, age. This procedure enabled us to explore the correlates of age-related variance and, furthermore, allowed examination of whether executive coordinative functions would provide further explanation of variance in prospective memory performance than did nonexecutive, rather basic storage functions. The results are summarized in Table 1. Verbal intelligence and retrospective memory did not significantly contribute to the prediction of vari TABLE 1 Hierarchical regression predicting prospective memory Predictors Step 1 Verbal intelligence Plan recall Step 2 Intention formation Stroop-inhibition Step 3 Age **p<.01.
β
R
2
∆R2
.02
.02
.35**
.33**
.45**
.10**
.09 –.09 29** –.14 –.43**
ance in executing prospective intentions (∆R2=.02). However, individual differences in planning and inhibition did explain a reliable amount of variance in carrying out the prospective memory task (∆R2=.33). By further including age, an additional amount of variance in the prospective measures was explained (∆R2 =.10). In sum, when all variables were considered in the regression, both planning (β=.29, p<.01) and age (p=–.43, p<.01) resulted as significant predictors of prospective performance.
36 KLIEGEL, MARTIN, MOOR
DISCUSSION The main objective of the present study was to examine the two variables task importance and cognitive resources/planning, both of which are believed to be related to delayed performance of complex multiple intentions. For this reason, prospective task importance was varied both during the planning phase as well as during the execution phase of the complex prospective multitask proposed by Kliegel et al. (2000). In addition, nonexecutive and executive cognitive resources were assessed in order to investigate the relationship between interindividual cognitive differences and age effects in complex prospective remembering. Overall, the results show age effects in favour of the younger group in all task components of the complex prospective multitask. In contrast, none of the groups profited significantly from the present experimental manipulation of importance. Finally, in regression analyses, planning (i.e., intention formation) in particular was found to be a significant predictor of intention execution, explaining most of the age-related variance. The present findings do not support the hypothesis that older people perform more poorly on a complex prospective task because the prospective task lacks importance. Remarkably, even combined hinting, i.e., emphasizing the necessity to interrupt/switch during the introduction phase as well as increasing task importance in the execution phase, led to no reliable improvements in older adults’ performance. Although post hoc questioning—in line with prior studies— revealed that the participants seriously intended to carry out all six tasks, explicitly increasing the amount of points possible for correct and complete switches between subtasks did not lead to significant improvements in performance. However, interpretation of this result is restricted in several ways. First, there are ceiling effects in the younger group’s data. While the issue of younger people’s performance regarding importance must therefore remain in the dark for the moment, older adults’ data suggest that such manipulations during the instruction phase and the execution phase do not, in fact, improve performance. Second, sample size was rather small and thus set a limit to interpreting data: Although it was big enough to detect large effects (Cohen, 1992), small or medium effects cannot be found with 12 participants per cell and p=.05. Consequently, interpretation must be qualified insofar as manipulating task importance has shown no large effect. Finally, it seems reasonable to argue that our manipulations may not have increased importance of switching, although that was their intention. At least with older people, indirect hinting by giving points may not have had the desired effect. In order to address this hypothesis, further studies examining the influence of importance should draw on stronger, verbally directed manipulations of task importance, as these have been used successfully with younger people (see Kliegel et al., 2001; Kliegel, Martin, McDaniel, & Einstein, in press; Kliegel & Martin, in press). Overall, though, the present findings seem
PROSPECTIVE MEMORY AND AGEING
37
to suggest that older adults do seriously intend to perform the delayed switches on their own, but then, actually, are not able to (or do not) do so when the moment arrives. In the main, the occurrence of age effects in the components of the complex prospective paradigm replicates the findings of Kliegel et al. (2000,2002) and Martin et al. (this issue; see also Einstein et al., 1992). Furthermore, the present study offers important clues as to how these age effects may have arisen: Consistent with our results, an explanation of the mechanisms of the reported age effects may be offered by assuming that executive functions, which are considered responsible for performance on complex tasks (Baddeley, 1996; Smith & Jonides, 1999) and which decrease with age (West, 1996), are related to older participants’ poorer delayed performance on complex multiple intentions (Kliegel et al., 2002; Marsh & Hicks, 1998; Martin & Schumann-Hengsteler, 2001; Maylor, 1996). Although correlated data do not allow us to infer causal relationships, converging with the paper by Martin et al. (this issue), regression analyses clearly underline that a large part of age-related variance in complex intention execution can be explained by executive measures. A result to further strengthen this conclusion is that nonexecutive processes, such as verbal intelligence or retrospective memory, do not—in contrast to prior findings (cf. Cherry & LeCompte, 1999; Einstein et al., 1992)—contribute to predicting variance in prospective performance, whereas both executive functions of planning and inhibition do predict prospective performance, even once nonexecutive variables are accounted for. Although age, considered in a third step, does provide an additional amount of explanation of variance (signalling that further processes not examined here may play a part), regression analyses suggest an important relationship between planning and inhibition processes on the one hand, and complex prospective remembering on the other, the latter causing difficulties—as analyses of means demonstrated—mainly for older adults. In sum, our results concerning performance of complex delayed intentions specifically highlight the fundamental importance of adequately planning the complex intention and argue against a motivational deficit in older adults. With support from several findings reported in planning literature (see Phillips et al., in press, for an overview), this finding provides empirical evidence for the multiprocess model of prospective remembering recently proposed by McDaniel and Einstein (2000), with planning of a delayed intention as one of the central processes. Consequently, one of the crucial factors in maintaining or even increasing independence in everyday life, for both younger and older adults, appears to be the sufficiently thorough planning of multiple future activities.
38 KLIEGEL, MARTIN, MOOR
REFERENCES Baddeley, A. (1996). Exploring the central executive. Quarterly Journal of Experimental Psychology , 49A , 5–28 . Bisiacchi, P.S. (1996). The neuropsychological approach in the study of prospective memory. In M.A. Brandimonte, G.O.Einstein, & M.A.McDaniel (Eds.), Prospective memory: Theory and applications (S. 297–317). Mahwah, NJ: Lawrence Erlbaum Associates Inc. Brandimonte, M.A., Einstein, G.O., & McDaniel, M.A. (Eds.) (1996). Prospective memory: Theory and applications . Mahwah, NJ: Lawrence Erlbaum Associates Inc. Burgess, P.W., Veitch, E., de Lacy Costello, A., & Shallice, T. (2000). The cognitive and neuroanatomical correlates of multitasking. Neuropsychologia , 38 , 848–863 . Cherry, K.E., & LeCompte, D.C. (1999). Age and individual differences in prospective memory. Psychology and Aging , 13 , 60–76 . Cohen, J. (1992). A power primer. Psychological Bulletin , 112 , 155–159 . Dempster, F.N. (1992). The rise and fall of the inhibitory mechanism: Toward a unified theory of cognitive development and aging. Developmental Review , 12 , 45–75 . Einstein, G.O., Holland, L.J., McDaniel, M.A., & Guynn, M. J. (1992). Age-related deficits in prospective memory: The influence of task complexity. Psychology and Aging , 7, 471–478 . Einstein, G.O., McDaniel, M.A., Manzi, M., Cochran B., & Baker, M. (2001). Prospective memory and aging: Forgetting intentions over short delays. Psychology and Aging , 15 , 671–683 . Ellis, J., & Kvavilashvili, L. (2000). Prospective memory in 2000: Past, present, and future directions. Applied Cognitive Psychology , 14 , 1-9 . Eschen, A., Kliegel, M., & Thöne-Otto, A. (2002). Neuropsychological mechanisms of prospective remembering . Manuscript submitted for publication. Hasher, L., & Zacks, R.T. (1988). Working memory, comprehension, and aging: A review and a new view. In G.H.Bower (Ed.), The psychology of learning and motivation (S. 193–225 ). San Diego, CA: Academic Press. Hohaus, L., Shum, D., & White, J. (2000). The effects of salience on prospective memory declines with age . Paper presented at the Cognitive Aging Conference, Atlanta. Houx, P.J., Jolles, J., & Vreeling, F.W. (1993). Stroop interference: Aging effects assessed with the Stroop colorwordtest . Experimental Aging Research , 19 , 209-224 . Kliegel, M. (in press). Gesundheitsverhalten bei chronischen Erkrankungen im höheren Erwachsenenalter. In A.Kruse & M.Martin (Eds.), Lehrbuch der Gerontologie: Alternsprozesse in multidisziplinärer Sicht . Bern: Huber. Kliegel, M., & Martin, M. (in press). Mechanisms underlying motivational effects in prospective remembering. In F. Columbus (Ed.), Advances in psychology research . Huntington, NY: Nova Science Publishers. Kliegel, M., Martin, M., McDaniel, M.A., & Einstein, G. O. (2001). Varying the importance of a prospective memory task: Differential effects across time- and eventbased prospective memory. Memory , 9 , 1-11 . Kliegel, M., Martin, M., McDaniel, M.A., & Einstein, G.O. (2002). Complex prospective memory and executive control of working memory: A process model. Psychologische Beiträge , 44 , 303–318 .
PROSPECTIVE MEMORY AND AGEING
39
Kliegel, M., Martin, M., McDaniel, M.A., & Einstein, G. O. (in press). Importance effects on performance in even-based prospective memory tasks. Memory . Kliegel, M, McDaniel, M.A., & Einstein, G.O. (2000). Plan formation, retention and execution in prospective memory: A new approach and age-related effects. Memory and Cognition , 28 , 1041-1049 . Kvavilashvili, L. (1987). Remembering intentions as a distinct form of memory. British Journal of Psychology , 78 , 507–518 . Kvavilashvili, L. (1992). Remembering intentions: A critical review of existing experimental paradigms. Applied Cognitive Psychology , 6 , 507–524 . Kvavilashvili, L., & Ellis, J. (1996). Varieties of intention: Some distinctions and classifications. In M.Brandimonte, G.O.Einstein, & M.A.McDaniel (Eds.), Prospective memory: Theory and applications (S. 23–51 ). Mahwah, NJ: Lawrence Erlbaum Associates Inc. Lehrl, S. (1977). Der Mehrfachwahlwortschatztest MWT - B. Erlangen: Straube. Marsh, R.L., & Hicks, J.L. (1998). Event-based prospective memory and executive control of working memory. Journal of Experimental Psychology: Learning, memory, and Cognition , 24 , 336–376 . Martin, M. (2001). Verfügbarkeit und Nutzung menschlicher Ressourcen im Alter: Kernpunkte eines gerontologischen Leitkonzeptes . Idstein: SchulzKirchner Verlag GmbH. Martin, M., & Kliegel, M. (2003). Die Entwicklung komplexer prospektiver Gedächtnisleistung im Kindesalter. Zeitschrift für Entwicklungspsychologie und Pädagogische Psychologie , 35 , 75–82 Martin, M., Kliegel, M., & McDaniel, M.A. (this issue). The involvement of executive functions in prospective memory performance of adults. International Journal of Psychology , 38 , 195–206 . Martin, M., & Schumann-Hengsteler, R. (2001). How task demands influence time-based prospective memory performance in young and older adults. International Journal of Behavioral Development , 25 , 386–391 . Maylor, E.A. (1996). Does prospective memory decline with age? In M.A.Brandimonte, G.O.Einstein & M.A. McDaniel (Eds.), Prospective memory: Theory and applications (S. 173–198 ). Mahwah, NJ: Lawrence Erlbaum Associates Inc. McDaniel, M.A., & Einstein, G. O. (2000). Strategic and automatic processes in prospective memory retrieval: A multiprocess framework. Applied Cognitive Psychology , 14 , 127–144 . McDaniel, M.A., Glisky, E.L., Rubin, S.R., Guynn, M.J., & Routhieaux, B.C. (1999). Prospective memory: A neuropsychological study. Neuropsychology , 13 , 103–110 . Phillips, L.H., MacLeod, M.S., & Kliegel., M. (in press). Adult aging and cognitive planning. In R.Morris & G. Ward (Eds.), The cognitive psychology of planning . Hove, UK: Psychology Press. Rendell, P.G., & Craik, F.I.M. (2000). Virtual week and actual week: Age-related differences in prospective memory. Applied Cognitive Psychology , 14 , S43–S62 . Salthouse, T.A. (1991). Mediation of adult age differences in cognition by reductions in working memory and speed of processing. Psychological Science , 2 , 179–183 . Shallice, T., & Burgess, P. (1991). Deficits in strategy application following frontal lobe damage in man. Brain , 114 , 727–741 . Smith, E.E., & Jonides, J. (1999). Storage and executive processes in the frontal lobes. Science , 283 , 1657–1661 .
40 KLIEGEL, MARTIN, MOOR
West, R.L. (1996). An application of prefrontal cortex function theory to cognitive aging. Psychological Bulletin , 120 , 272–292 .
The intention-superiority effect for naturally occurring activities: The role of intention accessibility in everyday prospective remembering in young and older adults Jayne E.Freeman and Judi A.Ellis University of Reading, UK This study examined age differences in the accessibility of a single pool of naturally occurring intentions both before and after completion. Following Maylor, Darby, and Della Sala (2000), accessibility was measured in terms of the number of activities generated in a 4-minute activity fluency task. Each participant undertook two such tasks. A prospective task in which they generated activities intended for completion during the following week and a retrospective task, 1 week later, in which they generated activities carried out over the previous week. In a partial replication of Maylor et al.’s findings, young, but not healthy older, adults generated more to-be-completed intentions than completed ones, demonstrating an intention-superiority effect (ISE) for everyday activities. The absence of an ISE for older adults appeared to reflect the reduced accessibility of intentions prior to completion, rather than the impaired inhibition of fulfilled intentions. Moreover, both groups showed greater inaccessibility of completed than intended activities, thus demonstrating an intentioncompletion effect for naturally occurring intentions that is preserved in healthy ageing (cf. Marsh, Hicks, & Bink, 1998). Despite showing a reduced accessibility of intended activities, older adults reported having completed a greater proportion of their intentions during the week than young adults. Moreover, there was a correlation between the ability to access intentions and the proportion of intentions completed only for young adults. These observations suggest that older adults’ everyday prospective memory performance may be relatively less dependent on intention accessibility and more dependent on other factors. While there was no age difference in the reported use and effectiveness of external retrieval aids, older adults demonstrated a greater level of temporal organization in the production of their intentions in the fluency task. This is consistent with the possibility that older adults may have more structured daily lives and may be able to use information about the sequence of
42 FREEMAN AND ELLIS
ongoing events to support superior everyday prospective remembering. Cette étude examinait les differences reliées a l’âge dans l’accessibilité d’un fond commun d’intentions se produisant de façon naturelle a la fois avant et après l’exécution. En accord avec Maylor, Darby, et Della Sala (2000), l’accessibilité fut mesurée en fonction du nombre d’activités générées dans une tâche d’aisance d’activité de 4 minutes. Chaque participant a pris part a deux tâches de ce genre: une tâche prospective, dans laquelle il a produit des activités devant être effectuées au cours de la semaine suivante, et une tâche retrospective exécutée une semaine plus tard, dans laquelle il a produit des activités menées durant la semaine précédente. Une reproduction partielle des résultats de Maylor et al. fut obtenue, avec les adultes plus jeunes, mais pas les adultes plus ages en santé, qui ont généré davantage d’intentions a être complétées que d’intentions complétées. Ces résultats démontrent un effet de supériorité de l’intention (ESI) pour les activités quotidiennes. L’absence de ESI pour les adultes plus ages semble refléter l’accessibilité réduite des intentions antérieurement a l’exécution plutôt que l’inhibition diminuée des intentions accomplies. De plus, les deux groupes ont montré davantage d’inaccessibilité des activités complétées que des activités prévues, ce qui démontre que l’effet de l’intention d’exécution, pour les intentions se produisant de façon naturelle, est preserve chez les individus plus ages en santé (voir Marsh, Hicks, & Bink, 1998). En dépit de l’accessibilité réduite des activités prévues, les adultes plus ages ont rapporté avoir complété une plus grande proportion de leurs intentions durant la semaine comparativement
© 2003 International Union of Psychological Science http:// DOI: 10.1080/ www.tandf.co.uk/ 00207590244000205 journals/pp/ 00207594.html
Requests for reprints should be addressed to Jayne Freeman, School of Psychology, University of Reading, Earley Gate, Whiteknights, Reading, RG6 6AL, UK (E-mail:
[email protected]). This work was supported by an Alzheimer’s Society research fellowship awarded to the first author, grant number QRD/99/ 04/02. Interim reports of the experiment were presented at the Third International Conference on Memory in Valencia, Spain, July 2001.
THE ROLE OF INTENTION ACCESSIBILITY
aux jeunesadultes. En outre, une corrélation fut soulevée entre l’habileté a accéder aux intentions et la proportion d’intentionscomplétées seulement pour les jeunes adultes. Ces observations suggèrent que la performance de la mémoire prospectivequotidienne des adultes plus ages peut être relativement moins dépendante de l’accessibilité de l’intention et plus dépendanted’autres facteurs. Tandis qu’aucune difference reliée a l’âge ne fut soulevée quant a l’utilisation et a l’efficacité des aides derecuperation externes rapportées, les adultes plus ages ont montré un niveau supérieur d’organisation temporelle dans laproduction de leurs intentions dans la tâche d’aisance. Ceci concorde avec la possibilité que les adultes plus ages peuventavoir une vie quotidienne plus structurée et peuvent être capables d’utiliser l’information concernant la sequence desévénements en cours pour supporter la supériorité du rappel prospectif de tous les jours. Este estudio examinó las diferencias de edad en la accesibilidad a un fondo común de intenciones que ocurren naturalmente tanto antes como después de terminar. De acuerdo con Maylor, Darby, y Della Sala (2000), la accesibilidad se midió en terminos del numero de actividades generadas en una tarea de fluidez de actividad de 4 minutos. Cada participante emprendio dos de esas tareas. Se pretendía que se terminara una tarea prospectiva en la que generaran actividades durante la siguiente semana, y una tarea retrospectiva, una semana después, en la que generaran actividades llevadas a cabo durante la semana anterior. En una réplica parcial de los hallazgos de Maylor et al., los adultos jóvenes, pero no los ancianos saludables, generaban más intenciones que debían terminar que intenciones terminadas, lo que demuestra un efecto de la superioridad de la intención (ESI) en el caso de las actividades cotidianas. La ausencia de ESI en los adultos mayores parecía reflejar la accesibilidad reducida de las intenciones antes de terminar, más que la inhibición disminuida de las intenciones logradas. Es más, ambos grupos mostraron mayor inaccesibilidad de actividades terminadas que de actividades intentadas lo que demuestra un efecto de terminación de la intención para las intenciones que ocurren naturalmente, lo que se conserva en el envejecimiento saludable (cf. Marsh, Hicks, & Bink, 1998). Independientemente de mostrar una accesibilidad reducida de actividades intentadas, los adultos mayores manifestaron haber terminado una mayor proporción de sus intenciones durante la semana que los adultos jóvenes. Es más, hubo una correlación entre la habilidad para realizar las intenciones y la proporcion de intenciones terminadas sólo en el caso de los adultos jóvenes. Estas observaciones sugieren que el desempeño de la memoria prospectiva cotidiana de los adultos mayores podría ser relativamente menos
43
44 FREEMAN AND ELLIS
dependiente de la accesibilidad de la intención y más dependiente de otros factores. En tanto que no hubo diferencia de edad en el uso de las ayudas externas de recuperación y en su efectividad, los adultos mayores demostraron un mayor nivel de organización temporal en la produccion de sus intenciones en la tarea de fluidez. Esto es consistente con la posibilidad de que los adultos mayores tengan vidas diarias más estructuradas y sean capaces de usar la información acerca de la secuencia de sucesos para apoyar un recuerdo prospectivo cotidiano superior. INTRODUCTION A key characteristic of most prospective memory tasks is the need to retain an intended action in memory until an appropriate context for its execution is encountered (e.g., passing a message to a friend tomorrow). A growing body of evidence suggests that information describing the details of a to-be-enacted action may have different representational properties during the retention interval (between intention formation and retrieval) from information that is not associated with an intention to act. This paper examines age differences in the representation of naturally occurring (or everyday) activities both prior to and following completion. The accessibility of intended activities The representation of intended activities was examined by Goschke and Kuhl (1993), who observed faster recognition latencies for words belonging to a “prospective” action script that participants were intending to execute later than for words belonging to a “neutral” script that was not designated for enactment (see also Freeman & Ellis, in press-a). This “intention-superiority effect” (ISE; Goschke & Kuhl, 1993) has since been replicated using lexical decision latency as an index of intention accessibility (Marsh, Hicks & Bink, 1998). The dynamic properties of intention-related information have been further clarified by the finding that lexical decision latencies are slower for prospective as compared with neutral items once the intended activity has been carried out (“intentioncompletion effect”; Marsh et al., 1998) or following cancellation of the intention (Marsh, Hicks, & Bryan, 1999). This apparent inhibition of intention-related information following successful completion or cancellation presumably functions to prevent an already satisfied or no longer necessary action being erroneously triggered each time the conditions for its execution are met. Recent research has also demonstrated that these properties of intended and completed/ cancelled intentions (i.e., heightened accessibility prior to completion and inhibition following cancellation) also hold for more naturalistic intentions carried out in the laboratory (Dockree & Ellis, 2001).
THE ROLE OF INTENTION ACCESSIBILITY
45
Surprisingly, given recent interest in the intention-superiority effect, there have been very few attempts to examine its role in successful prospective remembering (see Kuhl & Goschke, 1994, for an exception). Theoretically, there are a number of ways in which the heightened accessibility of an intention representation during the retention interval could influence the likelihood of the intended action being retrieved and executed at an appropriate moment. For example, action representations with a high level of accessibility relative to other representations in long-term memory are more likely to be recollected, either spontaneously in the course of other ongoing activities (e.g., Ellis & NimmoSmith, 1993; Kvavilashvili, 1987) or in response to a more strategic evaluation of the current status of one’s goals and intentions (Mäntylä, 1996). These recollections could increase the likelihood that the intention will be subsequently retrieved either by reviving its resting level of activation, or by prompting additional planning processes aimed at refining and consolidating the details of the action and its designated retrieval context (Ellis, 1996). Ageing and intention accessibility Given the potential importance of the ISE to successful prospective remembering, the question arises as to whether individuals who experience difficulties in carrying out their intentions do so, in part, because they fail to display or maintain a heightened level of accessibility for to-be-enacted actions during the retention interval. One possibility, for example, is that impaired intention accessibility could underlie age-related changes in prospective memory performance where these occur. Although findings regarding age differences in the ability to carry out intentions have been mixed, a number of studies have demonstrated that, at least on certain prospective memory tasks, older adults do perform less well than young adults (see Maylor, 1996). A recent investigation by Freeman and Ellis (in press-b), however, failed to observe any age differences in the size of the intention-superiority effect for laboratory-based activities. Young and older adults were asked to learn lists of simple actions (e.g., sweep, chop) for either verbal report or enactment during a subsequent recall test. Method of encoding was also manipulated so that lists were either enacted during encoding (subject-performed tasks or SPTs; Cohen, 1981) or verbally encoded (read only). The two age groups showed an equivalent increase in accessibility (i.e., faster latencies in an intervening recognition test) for both enacted and to-be-enacted material as compared with material that was verbally encoded for verbal report. This finding of preserved intention-superiority effects for healthy older adults contrasts with the results of a study by Maylor, Darby, and Della Sala (2000) examining age differences in the accessibility of naturally occurring intentions (i.e., intentions that an individual forms as part of his/her everyday life). In a speeded fluency task, young and healthy older adults and Alzheimer’s disease patients were asked to generate both a list of activities that they were intending to
46 FREEMAN AND ELLIS
carry out over the following few days/week and a list of activities they had completed over the past few days/ week. Four minutes were allowed for each task. Although young adults showed a tendency to generate a greater number of intended (prospective) activities than completed (retrospective) ones, demonstrating a form of intention superiority for naturally occurring intentions, the opposite was found for the older adults. That is, both healthy older adults and Alzheimer’s disease patients revealed an intention-inferiority effect insofar as they generated a greater number of their already completed intentions than intentions they had yet to fulfil. The failure of healthy older adults to demonstrate the increased accessibility for naturally occurring intentions observed in young adults is somewhat surprising given Freeman and Ellis’ (in press-b) findings using laboratory-based actions. It should be noted, however, that differences in age-associated declines in performance on laboratory and naturalistic tasks are not unprecedented (see Rendell & Craik, 2000, for discussion of this issue). This discrepancy is typically such that while age-related deficits are often apparent when participants are tested in a laboratory setting, older adults often show superior prospective memory performance when more naturalistic or real-world tasks are used. In other words, they outperform young adults on tasks that require them to carry out intentions in the context of their everyday lives. This is particularly interesting in light of Maylor et al.’s results, because it suggests that even though older adults may not show the same pattern of activation for naturally occurring intentions that is seen in young adults (i.e., higher accessibility for intended than completed activities), they may be more likely than young adults to carry out these intentions. As Maylor et al. (2000) acknowledge, the exact basis of age differences in the dynamic properties of naturally occurring intentions remains to be established. The intention-superiority and intention-inferiority effects observed in their study were ascertained by comparing the accessibility of intentions prior to completion with the accessibility of intentions that had already been carried out. This approach is somewhat different to that used in most laboratory-based studies in which the relative accessibility of intended and completed intentions is evaluated by comparing them both to non- intention-related (i.e., neutral) material (e.g., Marsh et al, 1998). Older adults have been found to perform worse than young adults on many tests of inhibitory function (e.g., McDowd, 1997). Thus, it might be expected that the intention-inferiority effect observed in Maylor et al.’s (2000) older adults (i.e., a tendency to produce more activities in the retrospective than the prospective fluency task) occurs because of an age-associated deficit in the ability to initiate the necessary inhibitory processes following intention completion. However, the absence of a neutral comparison condition in Maylor et al.’s paradigm makes it difficult to establish whether the advantage of intended over completed activities found for young adults is due to an increased accessibility of intentions prior to completion, a reduced accessibility (or
THE ROLE OF INTENTION ACCESSIBILITY
47
inhibition) of intentions following completion, or a combination of these. Accordingly, it is also not apparent whether observed age differences in performance on the two fluency tasks reflects an age-related decline in the accessibility of intentions during the retention interval or the impaired inhibition of intentions that have already been fulfilled. The purpose of the present study is to address this question by using a modified version of Maylor et al.'s task that will allow a more direct analysis of age effects on the relative accessibility of naturally occurring intentions both during and after the retention interval. Following Maylor et al. (2000), the current study used performance on speeded activity fluency tasks as a measure of the ease with which intention representations can be accessed at various stages of completion. Accordingly, the accessibility of intentions before and after completion was measured in terms of the number of items generated in each of the prospective (intended) and retrospective (completed) activity fluency tasks. Maylor et al.’s procedure was modified here, however, by asking participants to attend two sessions 1 week apart. In one (prospective) session participants were required to generate a list of intentions they planned to carry out over the following week, while in a second (retrospective) session they generated a list of all the activities they had carried out over the previous week. Thus the accessibility of a single pool of intentions was examined both during the retention interval prior to completion and at the end of the period in which they should have been performed. This allowed an examination of age differences in the proportion of intentions that were accessible in only one session or the other, i.e., those that were accessible prior to but not following completion, or vice versa. By testing participants at the beginning and end of a 7-day period, moreover, we were able to obtain information about the proportion of intended activities that were successfully completed in that period. This permits an exploration of how age differences in the ability to access intention representations during the retention interval translate into age differences in the ability to carry out intentions at the correct moment. It is hoped that this may help to resolve some of the noted discrepancies between laboratory-based and naturalistic studies in this area. METHOD Participants and design Sixty young adults aged 18–31 years (M= 20.95, SD = 3.38) and 60 older adults aged 60–83 years (M= 69.40, SD=5.57) were tested. Young participants were students from various departments at the University of Reading who received either course credit or a small payment ($5) for taking part. Older adults were all living independently in the community and were recruited via advertisements placed in local newspapers and community centres. They received travel costs and a small payment ($5).
48 FREEMAN AND ELLIS
Older adults scored more highly than young adults on the Vocabulary subtest of the Wechsler Adult Intelligence Scale-Revised (WAIS-R; Wechsler, 1981); t (118)=–4.27, p<.001 (young adult range=34–69, M =49.32, SD=6.89; older adult range=28–68, M= 55.30, SD=8.40). However, young adults had spent longer in full-time education than the older group; t(118)=4.98, p<.001 (young adult range=13–19 years, M=15.08, SD=1.42; older adult range=6–19, M= 12.50, SD=3.76). The main aim of the study was to examine age differences in fluency for intended (prospective) and completed (retrospective) activities. However, we also took the opportunity to elicit additional information about the characteristics of participants’ naturally occurring intentions. This was done in two ways. In one version of the study (Version 1), participants were asked to provide further details about their listed activities at the end of each session. In a second version (Version 2) participants were asked for additional information about their activities only at the end of the Session 2. This manipulation allowed us to take account of the possibility that answering questions about intended activities in the first session might influence participants’ subsequent performance in the second session or even the likelihood that they would carry out their intentions during the week. Twenty-six young and 29 older adults took part in Version 1 of the study, while 34 young and 31 older adults completed Version 2. Materials and procedure Participants were tested individually in two sessions that took place exactly 1 week apart. At the start of the first session all participants were given 4 minutes to generate a list of activities that they were planning to carry out over the course of the following week. Instructions were similar to those used in Maylor et al.’s (2000) study. Participants were told that they would be given a “Category Generation task” in which they would be presented with the name of a category and asked to generate as many items as possible from that category in a set amount of time. They were then given the following instruction: “The name of the category that I would like you to generate items from is ‘Jobs, appointments and activities that you intend to carry out over the next week’. I’d like you to spend the next 4 minutes giving me a list of all the items from this category that you can think of and I will write down everything that you tell me. Please give me just one or two words to describe each activity and please try not to mention routine activities that you would do automatically every day like having dinner or brushing your teeth.” This latter instruction was not part of Maylor et al’s original procedure, but was included here to prevent participants producing habitual activities rather than delayed intentions. The experimenter then started a stopwatch and wrote down all of the items generated by the participant. The task was stopped after 4 minutes. Following this prospective fluency task, participants in the Version 1 of the study were asked to provide further details about each of the activities they had
THE ROLE OF INTENTION ACCESSIBILITY
49
generated. Specifically, they were asked to state the time or occasion on which each activity should be performed and to rate the importance of carrying out the activity on a 5-point scale (1=not very important; 5=very important). They were also asked to rate the frequency with which they would normally perform each of the intended activities (1=less than once a month; 2 =once or twice a month; 3=once or twice a week; 4= about once a day; 5=several times a day) 1 . Participants in the second version of the study (Version 2) were not asked for any additional information at this point. On returning for the second (retrospective) session 1 week later, participants were again given general instructions about a Category Generation task before being informed that they would be required to spend 4 minutes producing a list of “Jobs, appointments and activities that you have carried out over the past week”. As in the first session, participants were asked to use a one- or two-word label for each item and to avoid generating any routine daily activities. No reference was made at this point to the fluency task that had been undertaken the previous week. The experimenter started timing the 4-min period and wrote down all activities generated. Following this retrospective fluency task, participants were shown the list of intended activities they had generated in the first session and were asked to place a tick next to all the activities they had successfully completed during the week. Participants were then asked to provide further details about the activities that they had produced for the first time in Session 2 (i.e., completed activities that had not been included in their list of intentions the previous week). For each of these new activities, it was of particular interest to ascertain when the intention to carry it out had first been created (i.e., whether the intention had been created prior to or following the first session). In other words, we wanted to establish which intentions had been available, but inaccessible, during the prospective fluency task in Session 1. This information was elicited in two ways. Participants in Version 1 of the study were asked to provide details about when they had first formed the intention to carry out the activity, while those in Version 2 were simply asked: “Was this activity planned more than one week ago (i.e., were you aware of the need to do it when you visited here last week)?” The type of retrieval aids (if any) used to support intention completion was also recorded for all activities, whether successfully completed or not. Finally, participants in Version 2 of the task (i.e., those providing no additional information about their intentions at the end of Session 1) were asked to give a rating of the frequency with which all of their activities, including those mentioned in the prospective fluency task but not completed, would typically be carried out.
1
‘Participants were asked a number of other questions about their intended activities at this point as part of a separate study on naturally occurring intentions. These data are not relevant to the current investigation of intention-superiority effects and so are not reported here.
50 FREEMAN AND ELLIS
Figure 1. Mean number (and standard error) of activities generated in the prospective and retrospective fluency tasks for young and older adults.
RESULTS As previously noted, it was considered possible that requiring participants to engage in an additional evaluation of their intentions in the prospective session in Version 1 could have modified their subsequent performance, both in the second session and throughout the intervening period. To address this, task Version (1 =prospective/retrospective questions vs. 2= retrospective questions only) was included as a between-subjects factor in all analyses (except where noted otherwise). For the sake of brevity, it is sufficient to note that there were no reliable main effects of task version in any of the following analyses and no reliable interactions involving this factor; all Fs<2, ps>.16. Individual statistics for the Version factor in each analysis, therefore, are not reported. Prospective and retrospective activity fluency The number of prospective (intended) and retrospective (completed) activities generated by young and older adults is displayed in Figure 1. A 2× 2×2 mixed ANOVA was conducted on these data, with Age (young, old) and Version (1, 2) as between-subjects factors, and Session (prospective, retrospective) as the within-subjects factor. There was a reliable main effect of session on activity fluency, F(1, 116)=10.28, MSe=11.40, p=.002, with a greater number of items being generated overall in the prospective than the retrospective session. The main effect of age on activity fluency was not reliable, F(1, 116)=2.04, MSe=46. 71, p=.156, however there was a reliable Session×Age interaction; F(1, 116)=19. 20, MSe=11.40, p<.001. Independent samples t-tests indicated that while older adults generated fewer intended activities than young adults in Session 1, t (118=18)=3.01, p=.003, there was no age difference in the fluency of retrospective activities as indicated by the number of items generated in Session 2; t<1. Paired t-tests performed on the data from each age group separately demonstrated that there was a difference in activity fluency between the two sessions only for young adults; t(59)=4.67, p<001. In keeping with the intention-
THE ROLE OF INTENTION ACCESSIBILITY
51
superiority effect demonstrated by Maylor et al. (2000), young adults generated a higher number of intended activities in Session 1 than completed activities in Session 2. There was no difference in fluency for prospective and retrospective activities in older adults; t<1. Therefore, although we have failed TABLE 1 Mean proportion (and standard deviation) of activities given a frequency rating of either 4 or 5 by young and older adults in the prospective and retrospective fluency tasks
Young adults Older adults
Prospective fluency task
Retrospective fluency task
.05 (.05) .03 (.05)
.04 (.06) .03 (.06)
to observe an advantage for intended over completed activities in older adults there is no evidence of the intention-inferiority effect for naturally occurring activities that Maylor et al. had observed in this age group. To overcome the possibility that participants might produce a list of habitual tasks (e.g., eating, dressing, etc.), as opposed to novel intentions, in the fluency tasks, the experimental instructions had included a request to avoid generating any activities that would be performed automatically on a daily basis. Some indication of the extent to which participants were able to adhere to this request can be gained by examining the frequency ratings that were given to items generated in each session. Specifically, the proportion of activities that were rated as occurring at least once a day (i.e., those with a frequency rating of either 4 or 5) was calculated and subjected to a 2×2×2 mixed ANOVA, with Age (young, older) and Version (1,2) as between-subjects factors and Session (prospective, retrospective) as a within-subjects factor. As can be seen from Table 1, the proportion of generated activities that were potentially habitual was very low overall (less than 4%). Moreover, this did not differ reliably between the two age groups or between the two sessions. There was also no reliable Age×Session interaction; all Fs < 3.3, ps > .05. This indicates that young and older adults were equally successful at avoiding the production of habitual tasks during the prospective and retrospective fluency tasks. Inaccessibility of prospective and retrospective activities While there was no indication of intention inferiority for older adults in the present study, this age group also failed to show any evidence of the intentionsuperiority effect that had been observed for young adults. The question therefore remains as to whether the absence of this advantage for to-becompleted intentions in older adults can be attributed to the impaired accessibility of intentions prior to completion, the impaired inhibition of intentions following completion, or a combination of these two deficits. In order
52 FREEMAN AND ELLIS
to examine these possibilities more closely, we classified each of the activities generated across the sessions into one of three categories: 1. Intended activities that were produced in the prospective session but not subsequently completed (I+C-items). 2. Completed activities, produced for the first time in Session 2, that the participant reported had not been existing intentions at the time of the prospective fluency task (i.e., the intention to carry them out was created after the first session) (I-C+). 3. Activities that were existing intentions at the time of Session 1 that were subsequently completed (I+C+). Importantly, it is possible to subdivide these I+C + items further into those that were produced in the fluency task in both sessions and those that were produced in one session only. This latter class of item is particularly interesting as it can be used to give an indication of the level of inaccessibility for intended and completed activities in each age group. If an item was produced for the first time in Session 2 but the participant subsequently reported that it had been established as an intention prior to Session 1 then this intention can be classified as having been available, but inaccessible, in the prospective fluency task. Similarly, if an activity is spontaneously produced only in the first session but the participant subsequently indicates that it was completed during the week (i.e., when re-presented with their prospective list in Session 2), then this item can be classified as being inaccessible in the retrospective fluency task, i.e., the item met the criteria for inclusion in the list but was not produced. The proportion of activities that were inaccessible in the prospective fluency task (i.e., inaccessibility of intended activities) was calculated as the mean proportion of intended and completed (I+C+) activities, produced in the retrospective fluency task, that had not been produced in the prospective task (i.e., items produced for the first time in Session 2 that the participant subsequently reported had been established as an intention prior to Session 1). The proportion of activities that were inaccessible in the retrospective fluency task (i.e., inaccessibility of completed activities), on the other hand, was calculated as the proportion of I+C+ activities that were produced in the prospective fluency task in Session 1 but not produced in Session 2 following their completion. By examining age differences in these two measures of inaccessibility we can evaluate whether the absence of an ISE for older adults can be best attributed to an impaired ability to access intended activities prior to completion or to the reduced ability to inhibit intended activities once they have been carried out. It is important to note that a simple comparison of the total number of items produced in the prospective and retrospective fluency tasks is not sufficient to address this question as the constitution of activities produced in these two tasks is potentially very different. For example, the prospective fluency task could include any activity that met the criterion of being intended for completion over the following week, regardless of whether it was subsequently completed or not
THE ROLE OF INTENTION ACCESSIBILITY
53
(i.e., a combination of I+C- and I+C+ items). The retrospective fluency task, on the other hand, could consist of any activity that had been carried out over the previous week. It could therefore include a number of activities for which an intention was formed before the start of the 7-day period (I+C+ items), as well as a number of activities that were carried out spontaneously during the week and thus would not have acquired the status of a delayed intention (I-C+). There may be important differences in the dynamic properties of these different sets of actions. However, equating for these potential differences by examining only those items that were completed during the week and for which an intention had been formed before Session 1 (I+C+) allows a purer measure of differences in the relative accessibility of activities at various stages of completion. The mean proportion of I+C+ activities that were inaccessible in each of the fluency tasks is displayed in Table 2. A 2×2×2 mixed ANOVA was performed on these data with Age (young, old) and study Version (1, 2) as between-subjects factors and Session (prospective, retrospective) as a within-subjects factor. There was a main effect of session on item inaccessibility, F(1, 116)=29.15, MSe=0.02, p<.001, with more activities being inaccessible in the retrospective than the prospective fluency task. It is not possible to ascertain from this finding whether these activities are inhibited relative to neutral (i.e., non-intention-related) material in memory. However, the finding of greater inaccessibility of I+C+ activities in the retrospective session than in the prospective session is consistent with the notion that intended activities undergo some form of inhibition following completion (cf. intention-completion effect; Marsh et al., 1998,1999). There was also a main effect of age on the inaccessibility of activities overall, with older adults producing more activities in one session that had not been produced in the other; F(1, 116)=6.03, MSe=0.08, p=.016. Importantly, there was also a reliable Age×Session TABLE 2 Mean proportion (and standard deviations) of activities that were inaccessible in each of the fluency tasks for young and older adults
Young adults Older adults
Inaccessibility of prospective items
Inaccessibility of retrospective items
.26 (.23) .40 (.25)
.42 (.23) .46 (.21)
interaction; F(1, 116)=5.83, MSe=0.02, p=.017. Independent samples t-tests indicated that older adults showed a higher level of inaccessibility for intended activities (prospective inaccessibility) than young adults; t(118)=3.24, p=.002. In contrast, however, there was no effect of age on the inaccessibility of retrospective activities; t(118) 18)=−1.00, p=.32. Paired samples t-tests performed on the data from each group separately indicated that there was a reliable difference in the inaccessibility of prospective and retrospective
54 FREEMAN AND ELLIS
activities for both young, t(59)=5.28, p<.001, and older adults, t(59)=12.61, p=. 035, with both groups showing a higher level of inaccessibility (inhibition) for intentions following than prior to completion. It therefore appears that the failure of older adults to generate a greater number of activities in the prospective than in the retrospective fluency task (i.e., the ISE demonstrated by young adults) reflects an age-related deficit in the ability to access intended activities rather than any deficit in the inhibition of already completed ones. Intention completion At the end of the retrospective fluency task in Session 2, participants had been represented with their original list of intentions from the previous session and asked to indicate which of the intended activities had been carried out. A 2 (Age) ×2 (Version) between-subjects ANOVA performed on the proportion of intentions completed during the week revealed a reliable age difference, F(1, 116)=7.62, MSe=0.02, p=.007, such that older adults reported having successfully completed a greater proportion of their intended activities (M= 0.80, SD=0.16) than young adults (M= 0.73, SD=0.14). This is in keeping with other studies demonstrating the superiority of older adults on naturalistic prospective memory tasks, but is somewhat surprising in view of the finding that older adults had a greater difficulty accessing their intended activities relative to young adults in the prospective fluency task. Unfortunately, we do not have any means of verifying participants’ reports of the particular intentions that they completed during the week. It therefore remains a possibility that older adults’ higher completion rate simply reflects an age-associated tendency to overestimate how well they have performed. This seems unlikely, however, in light of previous research on the accuracy of metamemory post-dictions, which demonstrates that young and older adults do not differ in their ability to judge how well they have performed on a prospective memory task (e.g., Kidder, Park, Hertzog, & Morrell, 1997). Previous studies have indicated that highly important intentions are more likely to be successfully completed than intentions associated with a lower level of TABLE 3 Mean importance ratings (and standard deviations) of completed and noncompleted intentions for young and older adults
Young adults Older adults
Completed intentions
Noncompleted intentions
3.87 (0.62) 4.06 (0.59)
3.27 (0.70) 3.55 (0.90)
THE ROLE OF INTENTION ACCESSIBILITY
55
importance (e.g., Maylor et al., 2000). One possibility, therefore, is that the observed age difference in completion rate reflects a tendency of older adults to attach a higher overall level of importance to their generated intentions than do young adults. This was investigated by conducting a 2×2 mixed ANOVA on the importance ratings from participants in Version 1 of the task (who gave ratings at the end of Session 1), with Age (young, older) as a between-subjects factor and Completion status (completed, noncompleted) as a within-subjects factor (see Table 3). Unsurprisingly, completed intentions had been given higher importance ratings in Session 1 than intentions that were not subsequently completed, F(1, 48)=23.88, MSe=0.32, p <.001. However, there was no reliable age difference in the rated importance of intended activities and no Age ×Completion interaction; both Fs <2, ps >.15, suggesting that overall differences in intention importance can not readily account for age differences in the proportion of intentions that are reported as having been completed. Taken together, these findings indicate that older adults report having carried out a greater proportion of their intentions during the course of the week than young adults, despite showing a reduced ability to access their intentions in the prospective fluency task. This suggests that there may be little correspondence between the ability to access one’s intentions in a speeded fluency task during the retention interval and the ability to subsequently retrieve and execute one’s intentions when an appropriate retrieval context arrives. To investigate this, the relationship between proportion of intentions reported as completed and prospective inaccessibility (as indicated by the proportion of intended and completed activities that were inaccessible in Session 1) was analysed. Separate Pearson’s correlational analyses were performed for each age group as it was considered possible that the relationship between these two variables might be different for young and older adults. For young adults, there was a reliable negative correlation between the inaccessibility of intentions in Session 1 and the proportion of intentions reported as completed, r(60)=–.363, p=.004, suggesting that young adults who were less able to access their intended (and subsequently completed) intentions in the prospective fluency task were also completing fewer of their intended activities throughout the course of the week. In contrast, however, there was no relationship for older adults between the inability to access intentions in the prospective session and the ability to carry out intentions, r(60)=−.124, p=.35. In other words, those older adults who demonstrated a high inaccessibility for intended activities in the prospective fluency task in Session 1 were no more likely to fail to execute their intentions than individuals with a relatively low inaccessibility of intentions. However, some caution should probably be used when examining the relation between these two measures, as they are based on two very different sets of items. While the completion measure is calculated only for those intentions that were accessible in the prospective fluency task in Session 1, the prospective inaccessibility measure is necessarily based on those items that were not accessible at this time. Nevertheless, the finding that the
56 FREEMAN AND ELLIS
relation between these two measures is markedly different for young and older adults is intriguing and will be considered in more detail later. The above analyses suggest that older adults may fail to show an ISE for naturally occurring intentions because they have an impaired ability to access their intentions prior to completion. That is, they have a greater proportion of I+C+ intentions that were inaccessible in the prospective fluency task than young adults. Despite this, however, older adults reported having completed a greater proportion of their generated intentions than did young adults. Moreover, despite finding a relationship in young adults between the ability to access to-becompleted intentions in the prospective fluency task and the ability to carry out intentions, there was no relation between these measures for older adults. Taken together, these findings suggest that older adults’ (superior) prospective memory performance might be less dependent than young adults’ on the activational status of the internal representation of the intended action and more supported by other factors. Use of retrieval aids One possible reason why older adults may have carried out a greater proportion of their intentions than young adults, despite showing reduced intention accessibility, could be that they are able to offset any effects of cognitive deterioration by using more effective retrieval aids. Participants had been asked at the end of Session 2 to provide information about any retrieval aids they had used to help them remember to carry out their intended activities. Most of the retrieval aids reported could be classified as external cues, including written reminders (e.g., a list or diary) and placing relevant objects in a prominent position. Of all the retrieval aids mentioned, approximately 92% fell into this category (90% for young adults and 95% for older adults). The proportion of intentions for which external retrieval cues were used was surprisingly small overall, and a 2 (Age)×2 (Version) between-subjects ANOVA indicated that this did not differ between young and older adults (M=0.24, SD=0.20 for young adults; M= 0.28, SD=0.20 for older adults); F<1. There was also no age difference in the apparent effectiveness of these cues, with young and older adults completing a similar proportion of the intentions for which an external cue had been used (M=0.89, SD=0.16 for young adults; M=0.86, SD=0.23 for older adults); F<1. It is arguably difficult to undertake a meaningful analysis of the completion rate of intentions for which non-external retrieval aids were used, as participants reported a highly diverse range of other strategies for these tasks. They included, for example, internal strategies such as rehearsal or imagery, as well as other more passive cognitive approaches (e.g., relying on someone else to remind them to perform the activity at the appropriate moment). It was of interest, however, to examine age differences in the completion of intentions for which participants had reported that no memory aids had been used. A 2 (Age)×2 (Version)
THE ROLE OF INTENTION ACCESSIBILITY
57
between-subjects ANOVA performed on these data indicated that older adults completed a reliably higher proportion of non-aided intentions (M= 0.77, SD=0. 21) than young adults (M= 0.68, SD=0.19); F(1,115)=7.62, MSe=0.04, p=.007. Taken together, these findings suggest that older adults’ ability to complete more of their intentions than young adults cannot be readily attributed to the greater use of external retrieval aids. Conversely, it appears to reflect the superior performance of older adults on delayed intention tasks for which no retrieval aid (external or otherwise) was used. Prospective fluency and temporal organisation A commonly invoked explanation of older adults’ superior performance on everyday prospective memory tasks (e.g., Rendell & Craik, 2000) rests on the assumption that they tend to have more structured daily lives than young adults. Older adults may therefore have more information available about ongoing sequences of events, which can be used to support intention realisation by providing a framework into which non-routine activities can be integrated. Interestingly, in line with this suggestion, it was noted during the prospective fluency task that older, but not younger, adults would often make statements that indicated they were relying on some pre-existing temporal structure (i.e., an internal calendar) to generate their intentions. For example, they would often make comments along the lines of “OK that’s all I can think of for Wednesday. Now let me think about what I’ve got to do on Thursday”. In some cases older participants explicitly told the experimenter that they were going to perform the task by first listing the activities that they planned to carry out immediately after the experimental session and then thinking through the rest of the week in order. Given these observations, it was of interest to examine age differences in the extent to which the generation of intended activities in the prospective fluency task was determined by a strategy of temporal organisation. At the end of the first session, participants in Version 1 of the task had been asked to provide additional information about the occasion on which each of their intended activities should be carried out (i.e., the designated retrieval context for each intention). This information can be used to give an indication of the amount of correspondence between the order in which the events were expected to take place in everyday life and the order in which they had been produced in the prospective fluency task. An index of the degree of temporal organisation in young and older adults’ fluency task output was calculated. This involved ordering the events in terms of the day/time on which they were due to be carried out and comparing this to the order in which they had appeared in the participant’s list of intended activities. The extent to which a participant’s production of activities deviated from their actual temporal sequence was derived by calculating the number of occasions on which an activity was produced in the fluency task immediately after an event that did not immediately precede it in everyday life. The observed number of deviations was then calculated as a
58 FREEMAN AND ELLIS
proportion of the maximum number of occasions on which it was possible for the order of production to deviate from the actual sequence of events. (The maximum number of possible deviations is equivalent to the total number of activities minus 1, as the first activity is not preceded by any other activity from which a deviation can occur.) The proportional deviation score was then subtracted from 1 to provide a measure of the correspondence between the order in which the activities were to be carried out in everyday life and the order in which they had been generated in the prospective fluency task 2 . This can be represented by the following formula: Consider an example in which a participant generates 12 intentions, 5 of which are preceded in the list by an activity that was not due to precede it in the real temporal sequence of events. The observed number of deviations is therefore 5 while the number of possible deviations is 11. This yields a temporal correspondence score for this participant of .55:1–(5/11). While the majority of intentions generated by participants were associated with a specific time-based retrieval context, some activities did not meet this criterion. For example, some activities were associated with a very loose temporal period (e.g., to be performed sometime in the next few days) or were linked to either an event- or activity-based retrieval context (e.g., “return books” when next passing library or “go shopping” when run out of essentials). For the purpose of calculating the index of correspondence, these items were treated in the same way as temporal deviations as their appearance at a given point in the prospective fluency task suggests an interruption in any ongoing temporal organisation in the output of activities. The proportion of intentions generated without a specific temporal retrieval context was quite low overall and did not differ between young and older adults (M= 0.23, SD=0.18 for young adults; M=0.28, SD=0.25 for older adults); t<1. There was also no reliable difference in the proportion of ongoing temporal sequences that were interrupted by the production of a temporally unspecified event (M=0.06, SD=0.07 for young adults; M=0.17, SD=0.39 for older adults); t(41)=1.28, p=.21. An independent samples t-test conducted on the index of correspondence scores indicated a reliable difference in the amount of temporal organisation demonstrated by young and older adults, t(51)=–3.11, p =.003, with older adults’ prospective fluency output revealing a higher degree of temporal organisation (M =0.37, SD=0.31) than that of young adults (M=0.16, SD=0.13) 3 . The overall proportion of activities generated in an order that was consistent with the order in which they should be carried out was relatively low. However, the fact that older adults maintained a higher degree of temporal consistency in the generation of 2
It is important to note that this figure does not represent the amount of deviation that would be expected if the participant was not using any temporal organisation techniques, as even a random production of activities would include some events that were produced in a temporal sequence purely by chance. However, it does provide a useful index of temporal deviation that can be compared between young and older adults.
THE ROLE OF INTENTION ACCESSIBILITY
59
their intentions than young adults suggests that they may have been relying more on the use of information about their ongoing sequence of daily events to retrieve their intentions in this task. It is entirely possible, in line with this, that older adults may also use this information more often, or more effectively, to support the realisation of intentions in everyday life. DISCUSSION This study set out to examine age differences in the accessibility of naturally occurring delayed intentions (cf. Maylor et al., 2000). The relative accessibility of intention representations at various stages of completion was examined by asking participants to perform a speeded fluency task for a single pool of activities both during and after the retention interval. In a partial replication of Maylor et al.'s findings, young adults were able to generate more intended activities at the start of a 1 -week period than completed activities at the end of the week, demonstrating an intentionsuperiority effect (ISE; Goschke & Kuhl, 1993) for naturally occurring intentions. While there was no evidence of the intentioninferiority effect observed for Maylor et al.'s older adults (i.e., enhanced fluency for completed as compared with intended activities), there was also no evidence of intention superiority in this age group. Ageing and the accessibility of intended and completed activities By examining the proportion of activities that were generated in one, but not both, of the fluency tasks it was possible to address the question of whether the absence of an ISE for older adults can best be attributed to an age-related decline in the accessibility of intended activities or to changes in the inhibition of intentions following completion. Adopting the assumption that items produced within the 4-min period of a given fluency task have a higher level of accessibility than items not produced within this time, it follows that activities produced in the prospective fluency task but not in the retrospective task were more readily accessible during the retention interval, when they represented intentions to act in the future, than when they were associated with an already completed intention. Conversely, those generated in the retrospective, but not prospective, fluency task are assumed to have been more readily accessible when tagged as a completed intention than when they formed the basis of an intention that had yet to be fulfilled.
3
A similar result is obtained if the temporally unspecified events are removed from the computation of correspondence rather than being treated as temporal deviations (M=0.21, SD=0.14 for young adults; M=0.43, SD=0.32 for older adults); t(51)=–3.21, p<.01.
60 FREEMAN AND ELLIS
Age differences in the inhibition of fulfilled intentions (i.e., a reduced intention-completion effect for older adults) would be consistent with current views of cognitive ageing, which highlight the underlying role of impaired inhibitory function in many age-related deficits (e.g., Hasher & Zacks, 1988). There was, however, no apparent difference between young and older adults in the relative inaccessibility (inhibition) of completed intentions in the retrospective fluency task. That is, in both age groups an equivalent proportion of activities met the criteria for production in the retrospective fluency task but were only accessible in the prospective fluency task prior to completion (i.e., activities with a higher level of accessibility during the retention interval than after completion). Moreover, both groups demonstrated a higher level of inaccessibility for retrospective than prospective activities, providing evidence of an intentioncompletion effect for naturally occurring activities that is preserved in healthy ageing. The reduced accessibility (or inhibition) of intention representations following completion is presumably a functional mechanism by which perseveration or the unnecessary repetition of an already performed activity is avoided: A naturally occurring analogue of the intention-completion effect observed for laboratory-based tasks (Marsh et al., 1998, 1999). There was an age difference in the inaccessibility of intentions prior to completion, however, insofar as older adults had a greater proportion of intended activities that met the criteria for inclusion in the prospective fluency task but were not produced until the second session. In other words, while older adults were no more likely than young adults to produce completed intentions that had been inaccessible during the retention interval, they did have a reduced ability to access intention representations prior to execution; a greater number of these intentions were accessible only in the retrospective task. Given that young and older adults did not differ in their retrospective accessibility, the observed age difference is assumed to reflect an age-associated reduction in the accessibility of to-be-completed activities rather than the heightened accessibility (or reduced inhibition) of intentions following completion. Goschke and Kuhl (1993; see also Marsh et al., 1998) have argued that the intention-superiority effect for intended actions reflects an automatic tendency of goal-related information to be maintained at a heightened level of activation in long-term memory as compared with other information. One explanation, for example, is that intention representations acquire the status of source or goal nodes in long-term memory, which according to Anderson’s ACT* model (1983) would allow them to remain highly activated during the retention interval without the need for rehearsal or recoding processes. The current finding of an age-re-lated deficit in the accessibility of naturally occurring intentions is not consistent with the assumed automaticity of the intention-superiority effect. Instead the increased accessibility of to-be-completed intentions may depend, in part, on the initiation of attentionally demanding encoding or rehearsal operations that serve to establish and/or consolidate the intentional status of the action representation in memory (see also Altmann & Trafton, 2002). One
THE ROLE OF INTENTION ACCESSIBILITY
61
possibility, for example, is that strategic operations are necessary for activities to acquire a “commitment marker.” These markers may be explicitly tagged to an action representation during intention encoding and serve to “bias the retrieval process toward intention-related memory images” (Goschke & Kuhl, 1993). The finding that retrieval is less likely to be biased toward intention-related information in older adults than young adults therefore suggests an ageassociated deficit in the executive or attentional operations necessary for the acquisition and maintenance of this type of marker. This is consistent with neuropsychological evidence suggesting that successful intentional tagging of action representations may be dependent on the integrity of executive functions supported by the frontal cortex (Shallice & Burgess, 1991). Intention accessibility and intention completion The current findings (cf. Maylor et al., 2000) suggest that older adults may have a particular difficulty maintaining intention-related information in a heightened state of accessibility. Nevertheless, this group reported having successfully completed a higher proportion of their intended activities than did young adults. Moreover, there was no correlation between older adults’ prospective inaccessibility score and the proportion of intentions that they successfully completed. Older individuals with a higher proportion of inaccessible intentions in the prospective fluency task did not necessarily fail to carry out more of their intentions. This suggests that impaired intention accessibility may not be a key contributor to prospective memory failures in older adults where these occur. In contrast, there was a reliable negative correlation between the proportion of intentions that were inaccessible in the prospective fluency task and the proportion of intended activities carried out among young adults. While this evidence is only correlational and cannot therefore provide the basis of any interpretations about causality, it does provide preliminary support for the suggestion that the ability to maintain highly activated intention representations may be related to successful intention realization in some populations (i.e., young adults). Other studies have also demonstrated superior prospective memory performance for naturally occurring or naturalistic delayed intentions among older adults (e.g., Rendell & Craik, 2000). There are a number of potential reasons why this may be the case. One possibility is that older adults may rely more on the use of external retrieval aids to support intention realization (e.g., calendars) than young adults (e.g., Maylor, 1990). This was not found to be the case here, however, with both age groups reporting a similar use of memory aids. Moreover, there was no age difference in the proportion of intentions for which a cue had been used that were completed, indicating that older adults did not make use of retrieval aids either more frequently or more effectively than did young adults. There was an age difference, however, in the completion of intentions for which no retrieval aid had been used. Importantly, this was such
62 FREEMAN AND ELLIS
that older adults reported having completed a higher proportion of non-cued intentions than young adults. Ageing and temporal organization Another explanation for the dissociation between intention accessibility and intention completion in older adults can be derived from considering more closely the manner in which the prospective fluency task was undertaken by each age group. Specifically, there was evidence of an age-related increase in the degree of correspondence between the order in which intentions were generated in the task and the order in which they were intended for execution in everyday life. This suggests that older adults may have been relying more on information about the temporal organization and structure of events in daily life to facilitate intention retrieval, perhaps with the aim of overcoming the apparent reduction in the relative accessibility of these items in memory. The finding of enhanced temporal organization in the output of older adults’ responses in the prospective fluency task is consistent with the commonly held assumption that older adults tend to have more routine or highly structured daily lives than do young adults. If this assumption is correct, it follows that information about the orderliness of ongoing activities might be used to support older adults’ superior prospective memory performance for naturally occurring intentions, thus overcoming the effects of reduced intention accessibility (cf. Ellis, 1996). Previous evidence of the supportive role of daily structure in older adults’ prospective remembering has been reported by Maylor (1990). She found that older adults were more likely to remember to make a telephone call once a day for 5 days when they associated the activity with other routine daily events (so-called “conjunction” cues) than when internal or other external cues (e.g., diaries) were used. The exact role of daily structure in the fulfilment of delayed intention tasks in young and older adults remains to be established, however, particularly as Maylor’s study did not include a comparison of the use and effectiveness of conjunction cues between these two age groups. It is interesting to note in this regard that an attempt has been made to enhance older adults’ prospective memory performance in a laboratory setting by using tasks that are intended to mimic the richness and structure of daily life events (e.g., Rendell & Craik, 2000). Age-related declines have still been obtained under these conditions, however, perhaps because the tasks are not readily able to capture or recreate the familiarity and personal relevance of the individuals’ own routines. Intention-superiority effects for naturally occurring and laboratory activities The current findings reveal a clear age-associated impairment in the ability to access naturally occurring intentions in a speeded fluency task undertaken during the retention interval between intention formation and completion. This is in
THE ROLE OF INTENTION ACCESSIBILITY
63
contrast to the findings of Freeman and Ellis (in press-b), which demonstrated an equivalent advantage for to-be-enacted laboratory-based actions over actions not intended for enactment in young and healthy older adults. We have argued elsewhere (e.g., Freeman & Ellis, in press-a) that there may be similarities between the advantage for to-be-enacted laboratory-based actions and the advantage that is frequently observed for verbally presented action words that have been enacted during encoding (the subject-performed task effect; Cohen, 1981). More specifically, the intention-superiority effect for simple motor actions intended for enactment after a short delay might reflect the operation of covert motoric or SPT-type encoding or rehearsal operations aimed at preparing these actions for imminent execution. These could include operations for setting the parameters of the action schema to be executed in terms of its duration, direction, and force. The absence of an age difference in the accessibility of laboratory-based intentions mirrors the finding of reduced age-related declines in memory for SPTs and suggests that covert motoric processing may be undertaken relatively automatically for this type of material. The apparent discrepancy between age differences in the ISE for naturally occurring and experimental intentions might therefore reflect a fundamental difference in the nature of the activities involved in the two paradigms. Specifically, the exact motoric requirements of many naturally occurring intentions (e.g., “buy birthday present”) may not be sufficiently well specified at encoding (or throughout the retention interval), to allow the representation of these activities to benefit from the kind of preparatory processing that we have argued supports the representation of more well-defined (laboratory-based) actions. Indeed, not all naturally occurring intentions involve action-based responses. Some of the activities generated by participants in the prospective and retrospective fluency tasks, for example, could be classified as involving primarily verbal responses (e.g., to have a conversation with someone or to pass on a message), while others represent purely thought-based or cognitive tasks (e.g., “choose holiday destination”). The exact role of preparatory motoric processing in successful prospective remembering remains to be established, however, as laboratory studies of the ISE have typically used experimenterinitiated retrieval, which removes the need for participants to remember to carry out the actions for themselves when a designated retrieval context arrives. Conclusion In summary, this study revealed a clear age-related decline in the ability to access intention representations prior to completion, with more intended activities failing to come to mind in the prospective fluency task for older adults than for young adults. There was no apparent age difference in the inaccessibility (or inhibition) of already completed intentions, however, with both age groups demonstrating evidence of an intention-completion effect. Despite reduced intention accessibility during the retention interval, older adults reported having
64 FREEMAN AND ELLIS
carried out more of their intended activities during the week than did young adults. Interestingly, this appeared to be the case primarily for intentions for which no speciflc retrieval aids had been used. One possibility is that older adults may compensate for impaired intention accessibility by relying more on the ongoing sequence of daily routine events to support intention retrieval and execution. This is consistent with the observation of an age-related increase in the temporal organization of activities produced in the prospective fluency task. In line with this, while there was a correlation between intention accessibility and intention completion in young adults, suggesting a role for the intentionsuperiority effect in prospective memory performance in this population, there was no evidence of this relationship among older adults. REFERENCES Altmann, E.M., & Trafton, J.G. (2002). Memory for goals: An activation-based model. Cognitive Science , 26 , 39–83 . Anderson, J.R. (1983). The architecture of cognition . Cambridge, MA: Harvard University Press Cohen, R.L. (1981). On the generality of some memory laws. Scandinavian Journal of Psychology , 22 , 267–281 . Dockree, P.M, & Ellis, J.A. (2001). Forming and canceling intentions: Implications for prospective remembering. Memory and Cognition , 29 , 1139–1145 . Ellis, J.A. (1996). Prospective memory or the realisation of delayed intention: A conceptual framework for research. In M.Brandimonte, G.O.Einstein, & M.A.McDaniel (Eds.), Prospective memory: Theory and applications ( pp. 1–22 ). Mahwah, NJ: Lawrence Erlbaum Asssociates Inc. Ellis, J.A., & Nimmo-Smith, I. (1993). Recollecting naturally occurring intentions: A study of cognitive and affective factors. Memory , 1 , 107–126 . Freeman, J.E., & Ellis, J.A. (in press-a). The representation of delayed intentions: A prospective subject-performed task? Journal of Experimental Psychology: Learning, Memory and Cognition . Freeman, J.E., & Ellis, J.A. (in press-b). Aging and the accessibility of performed and tobe-performed actions. Aging, Neuropsychology and Cognition . Goschke, T., & Kuhl, J. (1993). Representation of intentions: Persisting activation in memory. Journal of Experimental Psychology: Learning, Memory and Cognition , 19 , 1211–1226 . Hasher, L., and Zacks, R.T. (1988). Working memory, comprehension and aging: A review and a new view. In G. H.Bower (Ed.). The psychology of learning and motivation , Vol 22 (pp. 193–225 ). San Diego: Academic Press. Kidder, D.P., Park, D.C, Hertzog, C., & Morrell, R.W. (1997). Prospective memory and aging: The effects of working memory and prospective memory task load. Aging, Neuropsychology and Cognition , 4 , 93–112 . Kuhl, J., & Goschke, T. (1994). State orientation and the activation and retrieval of intentions in memory. In J. Kuhl & J. Beckmann (Eds.), Volition and personality: Action versus state orientation ( pp. 127–153 ). Seattle, WA: Hogrefe & Huber.
THE ROLE OF INTENTION ACCESSIBILITY
65
Kvavilashvili, L. (1987). Remembering intention as a distinct form of memory. British Journal of Psychology , 78 , 507–518 . Mäntylä, T. (1996). Activating actions and interrupting intentions: Mechanisms of retrieval sensitisation in prospective memory. In M.Brandimonte, G.O.Einstein, & M.A.McDaniel (Eds.), Prospective memory: Theory and applications ( pp. 93–113 ). Mahwah, NJ: Lawrence Erlbaum Asssociates Inc. Marsh, R.L., Hicks, J.L., & Bink, M.L. (1998). The activation of completed, uncompleted and partially completed intentions. Journal of Experimental Psychology: Learning, Memory and Cognition , 24 , 350–361 . Marsh, R.L., Hicks, J.L., & Bryan, E.S. (1999). The activation of unrelated and canceled intentions. Memory and Cognition , 27 , 320–327 . Maylor, E.A. (1990). Age and prospective memory. The Quarterly Journal of Experimental Psychology , 42A , 471–493 . Maylor, E.A. (1996). Does prospective memory decline with age? In M.Brandimonte, G.O.Einstein, & M.A. McDaniel (Eds.), Prospective memory: Theory and applications ( pp. 173–197 ). Mahwah, NJ: Lawrence Erlbaum Asssociates Inc. Maylor, E.A., Darby, R.J., & Della Sala, S. (2000). Retrieval of performed vs. to-beperformed tasks: A naturalistic study of the intention-superiority effect in normal aging and dementia. Applied Cognitive Psychology , 14 , S83–S98 . McDowd, J.M. (1997). Inhibition in attention and aging. Journal of Gerontology: Psychological Sciences , 52B , P265–P273 . Rendell, P.G., & Craik, F.I.M. (2000). Virtual week and actual week: Age-related differences in prospective memory. Applied Cognitive Psychology , 14 , S43–S62 . Shallice, T., & Burgess, P. (1991). Higher-order cognitive impairments and frontal lobe lesions in man. In H.S. Levin, H.M.Eisenberg, & A.L.Benton (Eds.), Frontal lobe function and dysfunction ( pp. 125–138 ). New York: Oxford University Press. Wechsler, D. (1981). Manual for the Wechsler Adult Intelligence Scale-Revised . New York: Psychological Corporation.
66 FREEMAN AND ELLIS
Disentangling executive functions and memory processes in event-based prospective remembering after brain damage: A neuropsychological study Ute A.Kopp Humboldt-University, Berlin, Germany Angelika I.T.Thöne-Otto University of Leipzig, Germany Successful prospective remembering (PR) comprises at least two components: one retrospective component that refers to the encoding and retrieval of the content of the intention and a second prospective component that involves the retrieval of the intended action at the appropriate moment. Whereas the retrospective component is very similar to memory skills like learning and retention of new information, the prospective component is thought to rely mainly on executive function. Head injuries can disturb PR during both of these stages. To disentangle the relative impact of executive functions and retrospective memory processes on PR we embedded a PR task in a 2-back verbal working memory paradigm. Fifty-six neurological patients with brain damage of various aetiologies were divided preexperimentally into four groups on the basis of their delayed recall Index in the Wechsler Memory Scale-Revised (WMS-R, indicating absence or presence of deficits in retrospective memory functions) and their age-corrected score in the Behavioural Assessment of the Dysexecutive Syndrome (BADS, indicating absence or presence of deficits in executive functions). Additionally, 19 controls matched for age and education were examined. We found that patients with deficits in executive functions detected fewer cues than any other group irrespective of their retrospective memory performance. However, eight patients with severe anterograde memory deficits could retain neither the intention nor its content. Thus, intactness of the retrospective component seems to be a necessary but not sufficient prerequisite for successful prospective remembering. For the execution of the intended action itself executive functions play a critical role. Le rappel prospectif (RP) réussi comprend au moins deux éléments: une composante rétrospective qui fait référence a l’encodage et au retrait du contenu de l’intention et une seconde composante qui est prospective et qui implique le retrait, au moment approprié, de l’action projetée. Tandis que la composante
68 KOPP AND THÖNE-OTTO
retrospective est semblable aux habiletés mnémoniques comme l’apprentissage et la retention de nouvelles informations, la composante prospective est présumément reliée principalement a la fonction exécutive. Les dommages cérébraux peuvent nuire au RP à la fois en ce qui a trait à ces deux composantes. Pour clarifier l’impact relatif des processus des fonctions exécutives et de la mémoire retrospective sur le RP, nous avons inclu une tâche de RP dans un paradigme de mémoire de travail verbal. Dans cette tâche, une succession de lettres était présentée aux participants et ces derniers devaient comparer chacune des lettres observées avec l’avantdernière perçue. Cinquante-six patients neurologiques présentant des dommages cérébraux d’étiologies diverses furent divisés en quatre groupes avant l’expérimentation, sur la base de l’index du délai de rappel dans le Wechsler Memory Scale- Revised (WMS-R, indiquant l’absence ou la presence de deficit dans les fonctions de la mémoire retrospective) et du score corrigé pour l’âge dans le Behavioural Assessment of the Dysexecutive Syndrome (BADS, indiquant l’absence ou la presence de deficit dans les fonctions executives). De plus, 19 participants contrôles, dont l’âge et le niveau de scolarité correspondent a ceux des participants expérimentaux, furent également observes. Nous avons trouvé que les patients ayant des deficits des fonctions executives détectaient moins les signaux que n’importe quel des autres groupes, sans tenir compte de leur performance de mémoire retrospective. Cependant, huit patients présentant des deficits sévères de mémoire antérograde ne pouvaient pas retenir ni l’intention, ni le contenu. Par consequent, le fait que la composante retrospective soit intacte semble être un prérequis nécessaire mais non suffisant pour réussir le rappel prospectif. Pour l’exécution de l’action projetée elle-même, les fonctions executives jouent un rôle critique. © 2003 International Union of Psychological Science http:// DOI: 10.1080/ www.tandf.co.uk/ 00207590244000205 journals/pp/ 00207594.html
Requests for reprints should be addressed to Dipl.-Psych. U.A.Kopp, University Hospital Charité, Department of Neurology, Schumannstrasse 20/21, 10117 Berlin, Germany (E-mail:
[email protected]).
EVENT-BASED PROSPECTIVE REMEMBERING
69
E l recuerdo prospectivo exitoso (RP) comprende, por lo menos, dos componentes: un componente retrospectivo que se refiere a la decodificación y recuperación del contenido de la intención y un segundo componente prospectivo que incluye la recuperación de la acción intentada en un momento apropiado. En tanto que el componente retrospectivo es más similar a las habilidades de memoria tales como el aprendizaje y la retención de información nueva, se piensa que el componente prospectivo depende principalmente de la función ejecutiva. El daño cerebral puede perturbar el RP durante las dos etapas mencionadas. Con el propósito de desenredar el impacto relativo de las funciones ejecutivas y los procesos de memoria retrospectiva sobre el RP, se disfrazó una tarea RP en un paradigma de memoria de trabajo verbal de dos precedentes. Dentro de esa tarea se presentó una sucesión de letras a los participantes y éstos debían comparar cada letra que se presentaba con la anterior. 56 pacientes neurológicos con daño cerebral de diversa etiología se dividieron antes del experimento en cuatro grupos con base en su recuerdo retardado al responder a la Escala de Memoria Corregida de Wechsler (WMS-R, que indica la ausencia o presencia de deficiencias en las funciones de la memoria retrospectiva) y su calificación corregida para la edad en la prueba Behavioural Assessment of the Dysexecutive Syndrome (BADS, que indica ausencia o presencia de deficiencias de la función ejecutiva). Además, se examinó a 19 participantes de control apareados por su edad y educación. Se encontró que los pacientes con deficiencias en las funciones ejecutivas detectaban menos señales que cualquier otro grupo independientemente de como se desempeñara su memoria retrospectiva. Sin embargo, ocho pacientes cuya memoria anterógrada presentaba dificultades graves no pudieron retener ni la intención ni su contenido. Por consiguiente, parece ser que el que permanezca intacto el componente retrospectivo no es una condición suficiente para recordar en forma prospectiva con éxito. Las funciones ejecutivas desempeñan un papel crítico en la ejecución de la acción intentada. Prospective remembering (PR), the ability to execute previously scheduled activities as planned, is crucial for the management of everyday life (see ThöneOtto & Walther, this issue). Meeting a friend at 7 o’clock in the evening or dropping a letter into the letterbox on one’s way to work are just some examples of this kind of remembering. Typically, in the beginning of a PR task, the intention to do something, the intended action, and the retrieval context have to be encoded together. Subsequently, the intention must be held in memory over a more or less extended delay. When the retrieval context occurs one has to retrieve the intention and the corresponding action while inhibiting the ongoing
70 KOPP AND THÖNE-OTTO
activity in order to switch to and execute the intended action (Ellis, 1996). PR failures can occur at different stages during this process. On the one hand retention of the action or retrieval context may fail; on the other hand, retrieval of the action at the appropriate time or event may be missed. Einstein and McDaniel (1990) have termed these the retrospective and prospective components of a PR task. It is assumed that only the retrospective component (remembering that and what to do) is a “classical” memory function whereas the prospective component (noticing when to do something and initiating the appropriate action) presumably depends mainly on executive functions. Consistent with this view, Einstein, Holland, McDaniel, and Guynn (1992) found a correlation between target event memory and PR performance, and Hitch and Ferguson (1991) between the retrospective component of PR and measures of free recall. In a multitask PR paradigm Kliegel, McDaniel, and Einstein (2000) could show that the level of executive functioning (i.e., working memory and inhibition) predicted successful initiation and execution of a complex PR task (the prospective component) whereas retrospective memory did not. Studies with brain-injured patients also support evidence for a two-component model. Bisiacchi (1996) reported results of a study conducted by Sgaramella, Zattin, Bisiacchi, Verne, and Rago in 1993 in which two out of three patients with traumatic brain injury failed to remember to perform the PR task completely. This suggests that the patients did not forget to perform the required action at the right time but rather forgot the content of the intention itself. Sgaramella, Borgo, Fenzo, Garofalo, and Toso (2000) reported the results of six patients with Herpes Simplex Encephalitis in two PR tasks. All of them showed marked deficits in both tasks albeit for different reasons: one patient knew that there was an additional task but could not remember its content (failure of retrospective component), two patients could remember what they were supposed to do, but failed to do so during the task (failure of prospective component), and three patients had forgotten both components, the intention and its content (failure of retrospective and prospective component). All patients were impaired in retrospective memory; four of them had additional deficits in executive function. Thus, disentangling the impact of brain damage on the two components was restricted. Taken together, these results underline the theoretical notion of PR relying on both retrospective memory and executive functions. Yet so far only one study has addressed the relative contributions of each component to PR specifically. McDaniel, Glisky, Rubin, Guynn, and Routhieaux (1999) assigned older adults based on their so-called “hippocampal” and “frontal functioning” to one of four groups representing the four combinations of high versus low hippocampal and high versus low frontal functioning. Hippocampal and frontal functioning were assessed by two test batteries derived from a factor analysis of neuropsychological tests done by Glisky, Polster, and Routhieaux (1995), and hence were based on cognitive rather than anatomical characteristics. The participants had to perform an event-based laboratory prospective memory task
EVENT-BASED PROSPECTIVE REMEMBERING
71
that required them to indicate the appearance of predefined target words during the course of a multiple-choice general knowledge test. The authors reported superior PR performance for the high frontal functioning as compared to the low frontal functioning group. The hippocampal factor had no significant influence on PR. This study, for the first time, assessed the influence of executive functions and memory processes on PR directly. However, there were some constraints: First, even in the low functioning groups performance was in the normal range. Thus, retrospective memory may still be a limiting factor to PR once it is pathologically impaired, which is often the case after brain damage. Second, the composite scores were made up of several independent factor scores, each contributing with different weights to the theoretical constructs so that the impact of individual subtests on the composite scores remains unclear. Furthermore, the measures contributing to the hippocampal and frontal scores are only partly sensitive to retrospective memory loss or executive dysfunction. Accordingly, in the present study brain-damaged patients suffering from deficits of retrospective memory, executive functions, or both are examined. Analysing PR in subjects who are clearly impaired in their cognitive functioning might be a more stringent way to determine the relative contributions of retrospective memory and executive functions to PR. Although the retrospective component of PR is thought to be related to mediotemporal lobe memory systems and the prospective component to prefrontal lobe structures, for two reasons the present study is related solely to cognitive parameters. First, brain lesions are rarely restricted to a region of interest but often extend to adjacent areas. Second, and more important, patients often show a decline in cognitive function although the brain region supposed to sustain this function is obviously intact. For example, retrospective memory can be compromised without hippocampal damage when strategic retrieval capabilities are diminished (Moscovitch, 1994). Executive dysfunction, on the other hand, can be found not only with prefrontal lesions but also with temporal lobe damage or after diffuse subcortical lesions (Guthke, Walther, Matthes-vonCramon, & von Cramon, 2001). For these reasons, we decided to classify patients exclusively by cognitive parameters as our main interest was on the contribution of different functions, rather than brain structures, to PR. In this study we focused on an event-based PR where a more or less prominent environmental cue signalled that the intended action should be executed (e.g., putting a letter into the letterbox on your way home; Einstein & McDaniel, 1990). In contrast, time-based PR tasks require a self-initiated retrieval process of the intended action after or at a certain time (e.g., meeting a friend at 8 pm, taking the turkey out of the oven after 3 hours). As time-based tasks usually yield lower performance than event-based tasks (Shum, Valentine, & Cutmore, 1999), an event-based design was chosen to avoid ground effects in the most severely impaired patients.
72 KOPP AND THÖNE-OTTO
METHOD Participants and design Fifty-six outpatients of the Daycare Clinic for Cognitive Neurology at the University of Leipzig, Germany, and 19 healthy control persons, matched for age and education, participated in the study. Patients suffering from language disorders or visual field defects were precluded. Language and visual abilities were assessed in standard clinical procedures by experienced logopedics and orthoptists, respectively. Twenty-three of the patients had traumatic brain injuries, 16 had a cerebral vascular accident, 6 suffered from hypoxia, 3 had an intracerebral haemorrhage, and 2 a subarachnoidal haemorrhage. Two patients had undergone resection of cerebral tumours and six were classified as brain injuries with other aetiologies (see also Table 1). Nineteen controls were mached for age and education. Each subject signed an informed consent form and none of them had a history of psychiatric illness or substance abuse. The patients were categorized along two dimensions on the basis of their retrospective memory performance (RM) and executive functioning (EF). To assess RM performance patients completed the Wechsler Memory Scale-Revised (WMS-R; Wechsler, 1987). EF was assessed through the Behavioural Assessment of the Dysexecutive Syndrome (BADS; Wilson, Alderman, Burgess, Emslie, & Evans, 1996). The BADS assesses executive functioning by six subtests requiring simple and complex planning, estimation, problem solving, mental flexibility, and rule obedience in everydaytype tasks. As for the WMS-R, patients were categorized as impaired when the age-corrected delayed memory score was lower than 85 (1 SD below mean score). For the BADS the cut-off point was an age-corrected standard score of 90. This score is 5 points higher than 1 SD from mean. However, clinical observation of more than 400 patients in the Daycare Clinic of Cognitive Neurology had shown that persons with a score below 90 were already clearly impaired in everyday executive functioning. For the control group we did not expect pathological performance on any of the analysed measures. Of the 56 patients, 23 had BADS scores lower than 90, and 13 of them showed an additional impairment on the delayed WMS-R score (EF low-RM low group). Ten patients in the impaired executive functioning group reached delayed WMS-R scores above 85 (EF low-RM high group). Of the 33 patients with unimpaired executive functioning, 15 showed deficits in delayed recall (EF high-RM low group) whereas 18 subjects suffered neither from impaired executive functions nor from deficits in delayed recall (EF high-RM high group) (see Table 1). As also shown in Table 1, the patient groups did not differ with respect to the time since brain injury (F<1). Education and estimated premorbid IQ as assessed through a vocabulary test (Mehrfachwahl-Wortschatz-Intelligenztest; Lehrl, 1999) was comparable between groups. Age, however, varied significantly between groups, F(4, 70) =4.18, p<.01. Post hoc comparisons revealed that the
EVENT-BASED PROSPECTIVE REMEMBERING
73
EF low-RM high group was older than the EF low-RM low and the EF high-RM low group. TABLE 1 Descriptive statistics of experimental groups Executive functioning Low
Months since injury Age Educationa Premorbid IQb BADSc
High
RM low (n=13)
RM high (n=10)
RM low (n=15)
RM high (n=18)
32 (3/229)
21 (2/61)
17 (1/109)
15 (1/60)
33 (15/54) 10 (8/13) 101 (86/ 130) 74 (56/86)
55 (22/71) 11 (8/13) 104 (91/ 124) 80 (38/88)
38 (21/60) 10 (8/13) 102 (85/ 130) 103 (93/ 113) 70 (50/84)
40 (18/59) 11 (8/13) 114 (88/ 130) 107 (97/ 123) 107 (90/ 136)
Controls (n=19)
44 (18/68) 10 (8/13)
62 (48/82) 98 (89/117) WMS-R, DRd Data in parenthesis are minimum and maximum values. a Education=years of formal education. b Premorbid IQ=IQ estimated by a vocabulary test (Mehrfachwahl-WortschatzIntelligenztest, Lehrl, 1999). c BADS=age corrected standard score on the Behavioural Assessment of the Dysexecutive Syndrome (Wilson, Alderman, Burgess, Emslie & Evans, 1996). d WMS-R, DR=age-corrected delayed recall standard score on the Wechsler Memory Scale-Revised (Wechsler, 1987; German version Härting et al., 2000).
Materials and procedure The ongoing task in this study was a verbal working memory paradigm. Two hundred consonants, half in upper and half in lower case, were presented one at a time on a PC screen for 200 ms each. Interstimulus interval was 2500 ms. Subjects had to indicate by pressing one of two buttons on a keyboard whether the present stimulus was equal to that seen two screens before (the 2-back task). The PR task was to press a third button on the keyboard whenever an f, an n, or an m appeared, irrespective of its case. There were a total of eight cues, two in each block of 50 stimuli. Cue appearance was randomized. Prior to testing, participants were given oral and written instructions about the 2-back and the prospective remembering task. However, PR was not mentioned as primary study aim. Rather subjects were told that their ability to do several tasks in parallel would be assessed. As soon as the participants were able to repeat the instructions correctly, a training session comprising 50 letter stimuli and two prospective cues was conducted. If necessary, instructions and training
74 KOPP AND THÖNE-OTTO
Figure 1. Prospective remembering responses as a function of groups of participants.
session were repeated. Eight patients with severe anterograde memory deficits were not able to remember the PR task at all, so that they had to be excluded from the study. To obtain a retention interval prior to starting the 2-back task, a computerized operation span paradigm and a simple reaction time task, included for reasons not related to the present study, were administered. The two filler tasks required approximately 20 minutes; thereafter subjects started working on the 2-back task. Instructions concerning the PR task were not repeated at this time. After completion of the 2-back task subjects were asked if they remembered that there was an additional task and if they knew what it was. RESULTS Prospective memory task For the PR task the percentage of hits was computed for each subject (for descriptives see Table 2). None of the patients in the EF-low groups managed to respond to all 8 target letters. The controls, on the other hand, were the only group where nobody failed to answer to any target letter. The percentage of hits per subject (see Figure 1) were submitted to a 3×3 between-subjects ANOVA with the factors Executive Functions (low/high/ controls) and Retrospective Memory (low/high/controls). This analysis revealed a main effect of EF and PR performance, F(1, 70)=7.93, p<.01, whereas no such influence could be found for RM, F(1, 70)=2.45, p>.10. Also, there was no interaction between EF and RM, F(1, 70)=1.12, p>.20. Bonferroni corrected post hoc tests showed a better PR performance for the controls (p< .01) and the EF high groups (p=.01) as compared to the EF low groups. EF high groups and controls performed equally well (p=1.00). When asked about the task instructions at the end of the experiment, all subjects remembered that they were told to push a different button when certain letters appeared irrespective whether they were in capitals or not. Moreover, even those who had failed to respond to any target letter could recall at least one of them correctly. Because of the small number of items, target remembering was
EVENT-BASED PROSPECTIVE REMEMBERING
75
analysed by nonparametric Kruskal-Wallis tests. Both measures, EF and RM, turned out to have a significant effect on target remembering: EF X 2 =6.24, p< . 05; RM X 2 =6.94, p<.05. Post hoc Mann-Whitney U-tests showed that both impaired groups remembered fewer target items than the control group: EF low groups U=134.5, p<.05; RM low groups U=159.5, p< .05. All other comparisons lacked significance. Ongoing task Descriptive statistics for the ongoing 2-back working memory task are summarized in Table 2. The percentage of correct answers, missing reactions, and mean reaction time were analysed in separate ANOVAs with the betweensubjects factors Executive Functions TABLE 2 Means and mean standard error for the PR task and the ongoing task Variable
PR hits (%) 2-back correct answers (%)
2-back misses (%)
2-back reaction times (ms)
EF low-RM low (n=13) EF low-RM high (n=10) EF high-RM low (n=15) EF high-RM high (n=18) Controls (n=19)
24 (7.81)
69 (3.81)
8 (2.71)
1162 (130.86)
45 (8.17)
69 (2.94)
3 (0.80)
1129 (171.29)
55 (7.98)
72 (3.85)
7 (3.60)
1229 (83.92)
59 (7.13)
76 (3.50)
8 (3.85)
1104 (59.48)
64 (6.68)
76 (3.08)
8 (2.73)
1156 (81.15)
(low/high/controls) and Retrospective Memory (low/ high/controls). As the descriptives already suggested, the measures did not differ substantially between groups (all F values < 1). DISCUSSION This study aimed to disentangle the cognitive processes involved in prospective remembering by testing patients with specific cognitive deficits in an eventbased PR task. As a main finding, brain-damaged patients with impaired performance on neuropsychological tests of executive functions performed worse in the PR task than patients with unimpaired executive functioning. Retrospective memory performance had no influence on PR: High memory functioning patients and low memory functioning patients did not differ significantly from each other. At first sight this seems to imply that retrospective memory plays a negligible role in PR. However, nine patients suffering from severe anterograde memory
76 KOPP AND THÖNE-OTTO
deficits could not participate in the study because they forgot both the intentions and their contents through the course of the training session. Likewise, Alderman and Burgess (1993) reported a patient who—after herpes simplex encephalitis— suffered from severe amnesia and was unable to carry out any intention beyond the time limits of his RM abilities. Thus, intact retrospective memory seems to be a prerequisite for PR. The prospective component depends on successful encoding and retention of the intention during the preceding delay. However, once some basic RM skills are preserved, the impact of RM on PR seems to be relatively small. Accordingly, in the RM deficient groups memory skills did not influence PR performance in a significant way. Another explanation would be that the present paradigm required only low retrospective memory abilities. Indeed, rememembering three letters seems an easy task compared to the different subtests contributing to the WMS-R delayed recall score. Also, the demands on active retrieval once the cue was noticed were presumably small. Support for this assertion comes from a study by McDaniel, Guynn, and Einstein (1997), in which elderly adults detected over 90% of PR targets given that they were highly salient. This suggests that the to-be-executed action can be easily retrieved once the cue is noticed. Executive functions may be involved in processing and reaction to the cue in different ways. First, subjects may actively monitor the environment for cue appearance. In this case their behaviour would be comparable to a dual-task situation as cognitive resources have to be divided between performing the ongoing task and monitoring for the cue. The ongoing task in this study was a rather demanding one. Subjects had to indicate whether a letter presented on the screen was the same as the one presented two screens before. Such n-back tasks rely on executive processes like monitoring and working memory (Smith & Jonides, 1999). Thus, low EF patients could have been expected to perform worse in the ongoing task. Neither reaction time nor hits, however, differed from high EF patients (which also indicates that there were no differences in the tradeoff between speed and accuracy across groups). It therefore can be assumed that the low EF patients compensated for their executive deficits by allocating more resources to the ongoing task, thereby reducing the available resources for monitoring the cues. Another explanation refers to the activation level of the mental representation of the intention. Goschke and Kuhl (1996) suggested that delayed intentions are held in a higher activational state than other mental representations so that they can be retrieved more easily once the critical cue occurs. According to Einstein and McDaniel (1996), the cue first elicits a feeling of familiarity (noticing) that is followed by memory search for the content of the intention. As stated above, in the present paradigm the memory search component should be minor. Thus, noticing should depend on the ease with which the mental representation of the intention exceeds a given threshold and comes to mind. An ongoing task that relies heavily on executive control— like the one employed here—reduces PR performance considerably, whereas ongoing tasks that do not require EF leave
EVENT-BASED PROSPECTIVE REMEMBERING
77
PR performance unaffected (Marsh & Hicks, 1998). For the low EF groups this consumption of EF by the ongoing task may have been particularly harmful as it interfered with the feeling of noticing. Aetiology of the brain damage of patients with deficits in executive functions was rather heterogeneous and included diffuse brain damage as well as focal damages in various brain areas. Due to the lack of neuroimaging data we cannot speculate in detail about the brain areas involved in PR. However, it follows from our results that PR is not exclusively linked to a specified prefrontal brain region (i.e., BA 10; Burgess, Veitch, de Lacy Costello, & Shallice, 2000; Okuda et al., 1998). As prospective remembering is a complex construct, which not only relies on RM and EF but probably also on attentional processes, it is probably sustained by a widespread cortical network that is vunerable to lesions at different sites. Taken together, the present study showed that deficits in executive functioning as assessed by the BADS significantly reduced performance in an event-based prospective remembering task. This result supports earlier findings in healthy older adults (McDaniel et al., 1999). In addition, we found that minimally preserved retrospective memory is a necessary prerequisite for PR, an observation that only emerged in the most severely impaired patients. This underlines that the contribution of RM to PR is likely to be underestimated when only healthy subjects or patients with mild to moderate memory deficits are tested. The study was not designed to identify specific executive subprocesses crucial for PR but rather aimed to disentangle the meanings of EF and RM for prospective remembering. However, for future research it is important to examine which of the executive functions under discussion are involved in PR. For example, Kliegel et al. (2000) used a complex multitask PR paradigm that enabled the authors to disentangle planning, remembering, initiation, and completion of a PR task. This kind of task seems to be a promising way to examine the different phases of the PR process in neurological samples. Finally, as Burgess et al. (2000) demonstrated, it is possible to create designs allowing the assessment of PR performance via functional brain imaging (i.e., functional MRI). Those studies could help to clarify the changes of PR performance in brain-injured patients. REFERENCES Alderman, N., & Burgess, P.W. (1993). A comparison of treatment methods for behaviour disorder following herpes simplex encephalitis. Neuropsychological Rehabilitation , 4 , 31–48 . Bisiacchi, P.S. (1996). The neuropsychological approach in the study of prospective remembering. In M.Brandimonte, G.O.Einstein, & M.A.McDaniel (Eds.), Prospective memory: Theory and applications ( pp. 115–142 ). Mahwah, NJ: Lawrence Erlbaum Associates Inc.
78 KOPP AND THÖNE-OTTO
Burgess, P.W., Veitch, E., de Lacy Costello, A., & Shallice, T. (2000). The cognitive and neuroanatomical correlates of multitasking. Neuropsychologia , 38 , 848–863 . Einstein, G.O., Holland, L.J., McDaniel, M.A. & Guynn, M.J. (1992). Age-related deficits in prospective memory: The influence of task complexity. Psychology and Aging , 7, 471–478 . Einstein, G.O. & McDaniel, M.A. (1990). Normal ageing and prospective memory. Journal of Experimental Psychology: Learning, Memory, and Cognition, 4 , 717–726 . Einstein, G.O., & McDaniel, M.A. (1996). Retrieval processes in prospective memory: Theoretical approaches and some new empirical findings. In M.Brandimonte, G. O.Einstein, & M.A.McDaniel (Eds.), Prospective memory: Theory and applications ( pp. 115–142 ). Mahwah, NJ: Lawrence Erlbaum Associates Inc. Ellis, J. (1996). Prospective memory or the realisation of delayed intentions: A con ceptual framework for research. In M. Brandimonte, G.O.Einstein,& M.A.McDaniel (Eds.), Prospective memory: Theory and applications ( pp. 1–22 ). Mahwah, NJ: Lawrence Erlbaum Associates Inc. Glisky, E., Polster, M.R., & Routhieaux, B.C. (1995). Double dissociation between item and source memory. Neuropsychology , 9 , 229–235 . Goschke, T., & Kuhl, J. (1996). Remembering what to do: Explicit and implicit memory for intentions. In M. Brandimonte, G.O. Einstein, & M.A. McDaniel (Eds.), Prospective memory: Theory and applications ( pp. 53– 92 ). Mahwah, NJ: Lawrence Erlbaum Associates Inc. Guthke, T., Walther, K., Matthes-vonCramon, G., & von Cramon, D.Y. (2001). BADSEndlich ein Test zur Erfassung der Exekutivfunktionen . Paper presented at the 16th meeting of the Society for Neuropsychology, Leipzig, Germany. Härting, C., Markowitsch, H.J., Neufeld, U., Calabrese, P, Deisinger, K., & Kessler, J. (2000). Wechsler Gedächtnis Test-Revidierte Fassung (WMS-R) . Bern: Huber. Hitch, G.J., & Ferguson, J. (1991). Prospective memory for future intentions: Some comparisons with memory for past events. European Journal of Cognitive Psychology , 3 , 285–295 . Kliegel, M., McDaniel, M.A., & Einstein, G.O. (2000). Plan formation, retention, and execution in prospective memory: A new approach and age-related effects. Memory and Cognition , 28 , 1041–1049 . Lehrl, S. (1999). Mehrfachwahl-Wortschatz-Intelligenztest (4th ed.). Balingen: Spitta Verlag. Marsh, R.L., & Hicks, J.L. (1998). Event-based prospective memory and executive control of working memory. Journal of Experimental Psychology, Learning, Memory and Cognition , 24 , 336–349 . McDaniel, M.A., Glisky, E.L., Rubin, S.R., Guynn, M.J., & Routhieaux, B.C. (1999). Prospective memory: A neuropsychological study. Neuropsychology , 13 , 103–110 . McDaniel, M.A., Guynn, M.J., & Einstein, G.O. (1997). The role of frontal lobe functioning in temporally-based and cue-based prospective remembering in older adults . Paper presented at the Cognitive Aging Conference, Atlanta, GA. Moscovitch, M. (1994). Memory and working with memory: Evaluation of a component process model and comparisons with other models. In D.L. Schacter & E. Tulving (Eds.), Memory systems ( pp. 269–310 ). Cambridge, MA: MIT Press. Okuda, J., Fujii, T., Yamadori, A., Kawashima, R., Tsukiura, T., Fukatsu, R., Suzuki, K., Ito, M., & Fukuda, H. (1998). Participation of the prefrontal cortices in prospective
EVENT-BASED PROSPECTIVE REMEMBERING
79
memory: Evidence from a PET study in humans. Neuroscience Letters , 253 , 127– 130 . Sgaramella, T.M., Borgo, F., Fenzo, F., Garofalo, P, & Toso, V. (2000). Memory for/and execution of future intentions: Evidence from patients with Herpes Simplex Encephalitis. Brain and Cognition , 43 , 388–392 . Shum, D., Valentine, M., & Cutmore, T. (1999). Performance of individuals with severe long-term traumatic brain injury on time-, event-, and activity-based prospective memory tasks. Journal of Clinical and Experimental Neuropsychology , 21 , 49–58 . Smith, E.E., & Jonides, J. (1999). Storage and executive processes in the frontal lobes. Science , 283 , 1657–1661 . Wechsler, D. (1987). Wechsler Memory Scale-Revised . San Antonio, CA: The Psychological Corporation / Harcourt Brace Jovanovich. Wilson, B.A., Alderman, N., Burgess, P.W., Emslie, H., & Evans, J.J. (1996). Behavioural Assessment of the Dysexecutive Syndrome . Bury St. Edmunds, UK: Thames Valley Test Co.
80 KOPP AND THÖNE-OTTO
How to design an electronic memory aid for brain-injured patients: Considerations on the basis of a model of prospective memory Angelika I.T.Thöne-Otto and Katrin Walther University of Leipzig, Germany
There has been an enormous development in the field of electronic memory aids, such as mobile phones, organizers, or combinations of both. Nevertheless, these devices have hardly been useful in the neuropsychological therapy of memory-impaired patients so far. The model of prospective remembering by Ellis (1996) is used as theoretical background in order to (1) analyse patients’ needs concerning electronic memory aids, (2) evaluate the usefulness of electronic memory aids for patients described in the literature, as well as of two commercially available electronic memory aids, and (3) derive conclusions concerning demands on how to design a memory aid suitable for patients. While most of the devices designed for patients improve their reliability by compensating for lack of selfinitiated retrieval, most of them are not able to compensate for problems during execution or evaluation. Commercially available memory devices have the advantage of being available to everybody, but they can only support mildly impaired patients. Patients with more severe deficits, however, need a more suitable interface in order to facilitate data entry. In addition, interactive chains of actions should be implemented in order to support execution and evaluation of future intentions. MEMOS, a system in development that tries to realize these demands, is presented. I ly a eu un développement énorme dans le domaine des aidesmémoire électroniques, comme les téléphones mobiles, les organisateurs ou la combinaison des deux. Néanmoins, ces appareils ont a peine été utiles dans la thérapie neuropsychologique des patients ayant des troubles de mémoire. Le modèle de rappel prospectif de Ellis (1996) est utilisé comme base théorique afin (1) d’analyser les besoins des patients en matière d’aides-mémoire électroniques; (2) d’évaluer l’utilité des aides-mémoire électroniques pour les patients décrits dans la littérature, ainsi que de deux aidesmémoire électroniques disponibles sur le marché; (3) de tirer des conclusions concernant les demandes relatives au design souhaité
82 HOW TO DESIGN AN ELECTRONIC MEMORY AID
des aides-mémoire pour les patients. Tandis que la plupart des appareils destinés aux patients ont été améliorés en terme de fiabilité en compensant pour le manque de recuperation auto-initiée, la majeure part d’entre-eux n’est pas capable de compenser pour les problèmes se manifestant pendant l’exécution ou l’évaluation. Les appareils de mémoire disponibles sur le marché ont l’avantage d’être disponibles pour tout le monde, mais leur capacité a supporter les patients qui ont des facultés diminuées est limitée. Les patients avec des dificits plus sévères, cependant, ont besoin d’une interface plus convenable afin de faciliter l’entrée de données. De plus, les chaînes interactives d’actions doivent être implantées afin de supporter l’exécution et l'évaluation des intentions futures. MEMOS, un système en développement qui tente de répondre a ces demandes, est présenté. Ha habido un gran desarrollo del campo de las ayudas electrónicas a la memoria, tales como los teléfonos móviles, organizadores o las combinaciones de ambos. No obstante, hasta ahora, estos aparatos no han sido muy útiles en la terapia neuropsicológica de pacientes con problemas de memoria. Se utiliza el modelo de recuerdo prospectivo de Ellis (1996) como antecedente teórico para (1) analizar las necesidades de los pacientes respecto a las ayudas electrónicas, (2) evaluar la utilidad de las ayudas electrónicas para pacientes descritos en la literatura, así como de dos ayudas electrónicas disponibles comercialmente, y (3) derivar conclusiones respecto a los requerimientos sobre cómo diseñar ayudas electrónicas apropiadas para pacientes. En tanto que la mayoría de los aparatos diseñados para pacientes mejoran su fiabilidad alcompensar la falta de recuperación autoiniciada la mayor parte de éstos no pueden
© 2003 International Union of Psychological Science http:// DOI: 10.1080/ www.tandf.co.uk/ 00207590244000205 journals/pp/ 00207594.html
Requests for reprints should be addressed to Angelika I.T. ThöneOtto, University of Leipzig, Daycare Clinic of Cognitive Neurology, Liebigstr. 22a, 04103 Leipzig, Germany (E-mail:
[email protected]). The project was supported by a grant by Sächsische Ministerium für Wissenschaft und Kunst (SMWK) and is presently supported by the Bundesministerium für Bildung und Forschung (BMB+F).
238
83
compensar los problemas durante laejecución o la evaluación. Los aparatos disponibles comercialmente tienen la ventaja de estar disponibles para todos, perosólo pueden apoyar a pacientes con un deterioro ligero. Los pacientes con deficiencias más graves, sin embargo, necesitanuna interfaz más apropiada para facilitar la entrada de los datos. Además, se debe implementar cadenas interactivas deacciones para apoyar la ejecución y la evaluación de las intenciones futuras. Se presenta un sistema en desarrollo que intentalograr estos requerimientos, MEMOS. INTRODUCTION Talking about everyday memory functions is frequently synonymous with talking about prospective memory (PM). PM has been shown to be highly correlated with subjective memory reports, and tests of PM in comparison to traditional memory tests, e.g., the Auditive Verbal Learning Test (AVLT), seem to be even better indicators of functional memory capacitiy (Kinsella, Murtagh, Landry, & Homfray, 1996). The relevance of PM for the patient’s ability to live independently could be shown in a study by Thöne and Walther (2001). In a semistandardized interview, the authors asked 53 patients with brain injuries of different aetiologies about their social living conditions and their use of memory aids. After the interview subjects were classified into independently vs. dependently living patients according to predefined indicators of the ability to manage one’s financial affairs and to run one’s home independently. There was a relation between the severity of brain injury, use of compensatory strategies, and the degree of autonomy. The severity of memory deficits (measured with the Wechsler Memory Scale-Revised; Wechsler, 1987) was a good predictor of the ability to live independently. Results also demonstrated that it is clearly harder to master everyday life if more than one cognitive function is impaired. Independently living subjects used significantly more memory aids such as diaries, notebooks, or internal memory strategies (e.g., imagery or clustering of information). In contrast, dependently living subjects relied on their relatives as memory aids. The authors concluded that successful compensation for prospective memory deficits is a relevant predictor of living independently after brain injury. At least a minimum of cognitive resources, however, seem to be necessary for successful use of compensatory memory strategies. Defining patients’ problems on the basis of a model of PM (Ellis, 1996) Ellis (1996) distinguishes five relevant phases involved in the realization of a delayed intention: encoding, delay, performance interval, execution, and
84 HOW TO DESIGN AN ELECTRONIC MEMORY AID
evaluation. On the basis of both the literature concerning the cognitive functions involved and our clinical experience with patients, the model can be used to analyse patients’ problems in each of these phases. 1. Encoding: Patients may have problems with listening to instructions due to attentional problems; there may be problems of comprehension or they may listen but not be able to encode the information. 2. Delay: Even if information was sufficiently encoded, it may extinguish from memory during the delay. 3. Performance interval: Retrieval cues may not be salient enough for initiating retrieval, or self-initiated retrieval may not be possible in time-based tasks. 4. Execution: Patients may have forgotten what they wanted to do because of retrospective memory problems, or they may have problems with execution due to other neuropsychological deficits such as executive dysfunction or apathy. Even if execution was successfully started patients may be interrupted and may not find their way back to the task. 5. Evaluation: Postponing tasks may be necessary, if they could not be executed in the recall situation. This may be difficult for patients even if they managed all the other phases. Another potential problem during evaluation is that patients may not be able to extinguish the intention from memory even if it was successfully executed. In this case they may start again later (problem of perseveration). Defining demands on external memory aids on the basis of the model of PM The problem analysis allows us to define the requirements for compensatory external memory aids. 1. Encoding: The fact that patients log delayed intentions into a memory book or an organizer may support deep levels of processing. In the case of patients who are not able to log information reliably themselves, external aids may be operated by relatives and thus missing encoding may be compensated for. 2. Delay: The memory aid will store what is intended to be done, and when. 3. Performance interval: During the performance interval an alarm will go off and the aid will indicate what the patient wanted to do. A memory book in which a date is logged, but which does not initiate action at the right moment, may be of little help. Similarly an alarm that simply goes off, without indicating what has to be done, will not work. 4. Execution: During the execution phase, ideally patients should be guided through the different phases of an intention until it is successfully executed. 5. Evaluation: It must be easy to postpone intentions that cannot be executed at the time planned, and successful execution needs to be indicated as well. Thus at all time the patient has to be able to verify whether or not an intention was completed.
238
85
Electronic memory aids for patients As long as electronic memory aids exist, neuropsychologists have tried to use them for the therapy of memory impaired patients. Because commercially available electronic aids were usually too complicated to handle, patient adaptations were developed and evaluated. Hersh and Treadgold (1994) were pioneers in this field, with the development of NeuroPage. NeuroPage is a simple, portable paging system, which can be fastened to the patient’s belt. Appointments for the patient are entered into a computer database by a caregiver. At the critical time messages are recalled and passed by modem to a telephone company, which sends it to the pager and releases an alarm. On the display a message appears that tells the patient what to do. In a study with 15 patients Wilson, Evans, Emslie, and Malinek (1997) examined to what extent NeuroPage was able to improve the reliability of patients in the execution of delayed intentions. Their results showed that all patients profited from the pager and were able to fulfil significantly more intentions with the help of NeuroPage compared to a baseline without the device. In Great Britain, meanwhile, NeuroPage has received the status of a standard therapy in the treatment of memory-impaired patients (Wilson, 2002). Nevertheless, there are certain handicaps: Patients are not able to enter information themselves. This may be easily compensated for by caregivers during encoding. When it comes to execution and evaluation, however, patients are not able to indicate whether or not they were able to fulfil the indicated task. Therefore, they cannot postpone a task that could not be completely executed. Van den Broek, Downes, Johnson, Dayus, and Hilton (2000) used an electronic memory aid, which they called “Voice Organizer.” The Voice Organizer is a dictaphonelike instrument, which can be trained to recognize the patients’ voice. The patient can record appointments and intentions with date and time on the organizer, and at the critical time an alarm will automatically go off. The patient can play the reminding message as often as necessary. The entry of appointments is very easy, and patients can do it themselves. In a study with five patients all of them seemed to profit from the organizer. Kim, Burke, Dowds, Boone, and Park (2000) introduced a PSION palmtop computer. The instrument was given to 12 patients and they were trained in using it during speech therapy or occupational therapy. Subsequently they were supervised by regular phone calls in which they were asked if they still used the device and whether it was helpful. Nine of the 12 patients found the instrument useful, and 7 said they would like to continue using it after the study. One of the three patients who did not profit did not learn how to use it, in spite of intensive training. The other two did not see any necessity to use memory strategies. The authors do not report any quantitative data concerning usefulness but only patients’ opinions. In addition, they do not provide any details concerning training procedure.
86 HOW TO DESIGN AN ELECTRONIC MEMORY AID
Wade and Troy (2001) evaluated the use of standard mobile phones as memory aids. A computer software company was approached to help develop a computerized system to send reminder messages to standard mobile phones. Messages can be as long as necessary. Thus reminders can be tailored to the individual’s needs, and patients may be talked through subtasks and sequences. If the alarm is not answered, the computer will continue trying to send messages until it gets through. Messages can be specified as “high priority,” in which case unanswered calls will be transferred to caregivers, alerting them that the message was not received. Users can be asked for confirmation of task completion by simply keying their personalized number into the phone. This mechanism is supposed to increase autonomy and be reassuring for caregivers. All five cases described in the study profited from using the mobile phones. Subjects gained more independence and were able to fulfil tasks more reliably. Wright et al. (2001) adapted two styles of pocket computer memory aids for brain-injured patients. An interface, designed for the study, provided a diary with auditory alarms, a notebook, and links between diary entries and certain pages of the notebook. The authors were interested in the usefulness of data-entry tools. One style of notebook had an external keyboard, while the other was a palm-sized computer with a touch pad. Twelve patients (mean age 34) with traumatic brain injuries were given each of the computer memory aids for 2 months with a month break in between. They were trained to enter their everyday appointments and notes. They were given a manual—comprised mainly of annotated pictures of the various screens—which they could always consider during use of the device. After 3 weeks they were visited and were asked to demonstrate how to find, change, and enter details, as well as setting alarms and the use of the diary and notebook. All participants were able to use the computers and 83% found them useful. The frequency of usage, however, was very variable. There were no significant correlations between any single psychometric measure and either the number of total entries or new entries made. Only the correlation with the National Adult Reading Test (NART; Nelson, 1991), which measures premorbid intelligence, was marginally significant. More importantly, patients accustomed to using other memory aids before entering the study tended to make more use of the pocket computers. While high users preferred the computer with the physical keyboard, low users made more entries with the palmsized computer. The study was able to show that people with memory impairments may use individually designed computer-based memory aids. The design solution presented people with an unambiguous set of choices. It allowed them to rely on their problemsolving skills rather than expecting them to remember procedures for use. Table 1 gives an evaluation of the memory devices described in the literature on the basis of the model of PM by Ellis (1996). Several different electronic memory aids have been described in the literature so far. All the devices seem to increase subjects’ reliability, by compensating for their lack of self-initiated retrieval at the critical moment with an alarm and the reminding message. The question of whether the patients’ independence
238
87
increases seems to be highly related to the ease of data entry. Some of the devices can only be used by caregivers (e.g., NeuroPage), thus relatives may still be involved. On the other hand, other devices have no possibility of informing caregivers as to whether a task was carried out. If caregivers are used to looking after their brain-injured relatives and reminding them of relevant tasks, they may feel insecure about whether the system works reliably. Compensation for problems during the phases of execution and evaluation is only provided by few devices (e.g., the mobile phone system by Wade & Troy, 2001). This may be especially important when using external memory aids with severely impaired patients. Usefulness of commercially available electronic memory aids to compensate for PM problems in brain-injured patients The studies described so far use memory aids adapted to the patients’ needs. One problem of these instruments is that they are only available for patients taking part in the studies or affiliated with the research centres. In our own study (Guthke, Walther, & Thöne-Otto, 2002) we therefore wanted to know whether patients are able to use commercially available electronic memory aids, if they are trained to handle them. Therefore we examined the usefulness of a palm organizer (Palm ml00 00), as well as a mobile phone with agenda function. METHODS Participants Participants were 12 former male patients of the Daycare Clinic of Cognitive Neurology at the University of Leipzig. They had suffered brain injuries of various aetiologies. The patients were 28 to 58 years of age with a mean age of 46 years. All patients were at a stable, chronic stage, at least 28 months postinjuries (mean 46 months). The sample consisted of patients with unimpaired to moderately severe impaired memory functions. The subjects’ characteristics are presented in Table 2. Procedure An ABAC design was used to evaluate the utility of the two memory aids. During a baseline period of 2 weeks (Phase A) we asked patients to perform 20 experimental tasks without an electronic memory aid. The patients had to note forgotten intentions daily, send this note to the daycare clinic at the end of a week, and call a mailbox twice a week. At the same time, entering appointments into and receiving them from one of the two memory devices was trained.
88 HOW TO DESIGN AN ELECTRONIC MEMORY AID
Maximally five sessions of 60 minutes each were available to learn six different functions (e.g., entering appointments, postponing appointments, entering repeated intentions). Patients were given adapted instructions in order to support learning. After training, patients were given the memory device for the following 2 weeks (Phase B) and were asked to fill out a questionnaire on how useful they found the device. Again they had to perform the 20 experimental tasks, which had been entered into the memory aid by the experimenter before patients were given the device. After a break the same procedure was run with the second memory device (Phase A and C). The order of using the memory devices was counterbalanced across subjects. Indicators of success were: 1. progress during training (number of functions learned); 2. number of experimental tasks successfully executed during baseline and during use of the device; 3. number of everyday intentions forgotten during baseline and during use of the device; 4. number of everyday intentions entered into the device by the patient; and 5. patients’ satisfaction with the usefulness of the devices, measured using a questionnaire. TABLE I Evaluation ot electronic memory aids described in the literature on the basis of the model of prospective memory by Ellis (1996) Memory Aid
Encoding
Performance Execution interval
Evaluation
Comment
NeuroPage (Wilson et al., 1997)
Relatives do it.
An alarm goes off.
Patient is reminded once, no feedback.
Good as a reminder, but no feedback possible.
Voice Organizer (Van den Broek et al., 2000)
Appointme nts are spoken on the voice organizer.
Alarm goes off, instrument winds to the relevant point and plays the message as often as possible.
No feedback, no external control; patient can wind back by speaking the date and time, since the instrument can learn to recognize
No postponing of intentions possible, no indication whether intentions executed. Postponing is easy by entering a new date.
Good for fairly independent ly living subjects. No reassurance for caregivers. No guidance through actions.
238
Memory Aid
Encoding
PSION palm computer (Kim et al., 2000)
Patients enter appointmen ts. One subject was not able to learn it.
Mobile phone with computerize d system to send reminder messages (Wade & Troy, 2001)
From the description it is not clear whether patients can enter information themselves directly into the phone or if information needs to be entered into a computer base. Data entry with an external keyboard or the palm touch pad.
Pocket computer with and without external keyboard (Wright et al., 2001)
Performance Execution interval patient's voice. Alarm goes No external off, control. indicated message.
Evaluation
Comment
Patient may enter new appointmen t if necessary.
No quantitative data. No information concerning severity of memory deficit and training procedure. The system sounds very promising; details are not quite clear from the paper.
Patients receive a phone call that reminds them. The system makes sure that the message gets through or a caregiver can be informed.
According to the paper, messages could be as long as necessary and subjects could be talked through the different steps of a task. It is not clear how this works.
Patients can confirm execution by entering a PIN-code. Data-entry is not clearly described, so it is not clear whether patients can postpone appointmen ts.
Alarm goes off and shows message on the display.
Patient is reminded once, no feedback.
Patient can postpone intentions by entering them again.
A special design makes data entry easier. Good for fairly independent ly living subjects. No reassurance for caregivers. No guidance through actions.
89
90 HOW TO DESIGN AN ELECTRONIC MEMORY AID
TABLE 2 Demographic, medical, and neuropsychological information for each patient Patient
Age (years)
Aetiology
Time post onset (months)
Type of learner
WMS-R MQ index
1 53 SAH 32 S 65 2 50 TBI 73 S 69 3 38 TBI 86 S 78 4 37 other 73 S 78 5 58 TBI 29 S 86 6 38 TBI 45 S 88 7 56 CVA-r 28 F 78 8 38 other 29 F 84 9 28 TBI 63 F 84 10 31 TBI 64 F 94 11 53 CVA-r 34 F 100 12 48 SLE 38 F 118 CVA-r: cerebrovascular accident right; F: fast learner; S: slow learner; SAH: subarachnoid haemorrhage; TBI: Traumatic brain injury; SLE: systemic lupus erythematosus; other: other neurological disease.
RESULTS The results were related to severity of memory deficits. According to their progress during training, the patients could be divided into fast vs. slow learners. While slow learners (n=6) hardly learned the most basic functions of the devices, fast learners (n=6) had completed all six functions. The performance during training was highly correlated with retrospective memory measured by the WMSR (Wechsler, 1987). Both devices decreased the number of intentions forgotten. This effect was even higher for everyday intentions than for experimental tasks (see Figure 1). Looking at the spontaneous use of the devices in everyday life reveals a difference between fast and slow learners. Slow learners made less use of the devices, and entered more appointments into the mobile phone compared to the organizer (cf. Figure 2). Fast learners, in contrast, preferred to use the palm organizer. This result was confirmed by evaluation of the questionnaire data, in which fast learners were more likely to say that they would like to continue to use an electronic device. Asked for their preference, most of them favoured the palm organizer. Slow learners did not show a clear preference. On the basis of the questionnaire, several advantages and disadvantages of the two devices can be described. These are summarized in Table 3.
238
91
Figure 1. Mean number (and standard error) of delayed intentions forgotten during baseline and treatment. Experimental tasks were given by the examiners; everyday tasks were entered by the subjects.
Figure 2. Mean number (and standard error) of appointments entered by the patients as an indicator for device acceptance.
DISCUSSION In general, commercially available electronic memory aids may be an efficient compensatory strategy for patients with mild cognitive deficits. For those with more severe problems, specific adaptations need to be considered. On the basis of the PM model the following problems can be defined: 1. Encoding: For both devices the interface for entering appointments was too complicated for some patients. There were too many buttons and menu-steps, which often were not clearly labelled. Also, buttons were too small. 2. Performance interval: The alarm of the palm organizer was evaluated as too soft and too short. The display of the mobile phone was too small to give enough background information.
92 HOW TO DESIGN AN ELECTRONIC MEMORY AID
TABLE 3 Evaluation of user friendliness of two commercially available memory aids based on a questionnaire. Device characteristic Format (size, weight) Data entry
Palm organizer
Mobile phone
labelling of key is not clear Number of steps for entering information Keys
too many (19 steps)
too many (13 steps)
touch panel too small
keys too small, too many letters on one key
too small
too small
Letter size Memo text limited to 16 letters Alarm too soft and too short Presentation of missed intentions Overview of intentions only date and time indicates features that were evaluated as positive; indicates those that were criticized.
3. Execution: The memory devices do not offer the possibility of guiding patients through the different steps of an intention. Thus, if they were interrupted, they often did not find their way back to the task. 4. Evaluation: If an appointment had to be postponed, the entering problems mentioned for the encoding phase became relevant again. The devices do not ask automatically if an intention can be executed or if it needs to be postponed. Neither do they require confirmation of execution. Thus there is no feedback on whether an intention can be extinguished. Taken together, commercially available electronic memory aids may compensate for mild PM problems during encoding, delay, and performance interval. They are not suitable for more severely impaired patients and problems during execution and evaluation, due to neuropsychological deficits that cannot be compensated for with the available aids. How to design an electronic memory aid for brain-injured patients Based on the devices reported in the literature and the results from commercially available electronic memory aids, the Daycare Clinic of Cognitive Neurology and the Dept of Computer Science at the University of Leipzig collaborated to
238
93
design a memory system that is adaptive to patients’ needs. The system MEMOS (Thöne-Otto, Schulze, Irmsher, & von Cramon, 2001) consists of an internet server that allows the management of several clients, such as patients, therapists, and significant others. In addition there is an application server managing the execution of incoming and outgoing tasks. Appointments can be entered via different computers, which may be organized at a central service interface or in the patient’s home. In addition, appointments can be entered via speech input directly into the mobile device. Relevant patient data are stored in a database. From the application server, information is sent to the patient’ s personal memory assistant, PMA, a mobile memory device similar to a mobile phone. Interactive contact is possible between the application server and the PMA at any time. Thus the patient can be contacted directly if relevant appointments are not confirmed. In turn the patient can contact the service interface if further information is needed or in a case of emergency, in which the service interface may alert a carer or therapist. MEMOS is designed to take the aforementioned problems into account: 1. Encode: Encoding is very easily done, because patients can simply speak their appointments into the PMA. At the service interface, speech input is decoded by an assistant, in a similar way to listening to an answering machine, and data will be entered into the system via a PC. The PMA itself has only very few clearly labelled soft buttons. In the ideal case patients do not have to learn its use, but simply do what the device asks them to. 2. Delay: As with all the aids mentioned above, MEMOS is limited to timebased information. A future perspective may be to introduce GPS to locate the patient at any time and make demands on the basis of local information possible (e.g., if the tram reaches a certain stop the device will tell the patient to get off). 3. Performance interval: The alarm can be individually adapted. The display is large enough to give relevant information. Further background information can be ordered on request. 4. Execution: Patients are interactively guided through the different steps of an action. As far as possible this is programmed automatically, and each step has to be confirmed. On request, however, adaptations to the demands of an individual situation are possible. 5. Evaluation: Execution has to be confirmed. In the case of missing confirmation of relevant intentions (such as important medication), direct contact is possible. Patients are asked whether the task can be fulfilled or if it needs to be postponed. In the case of postponement, the system automatically looks for other appointments that may conflict with the postponed one. CONCLUSION Prospective memory is of high functional relevance in everyday life and the successful compensation for PM problems seems to be a decisive indicator of independent living. The model of PM can serve as a helpful background in order
94 HOW TO DESIGN AN ELECTRONIC MEMORY AID
to analyse patients’ needs and the demand for electronic memory aids. In this paper it was used to evaluate different electronic memory aids designed for patients as well as commercially available aids. On the basis of the model we were able to show that the aids so far mainly compensate for self-initiated retrieval. In order to help even severely impaired patients, however, further adaptations as outlined above seem to be necessary. During the development and design of the system MEMOS these considerations were taken into account. Evaluation of its practical usefulness is in progress. REFERENCES Ellis, J. (1996). Prospective memory or the realization of delayed intentions: A conceptual framework for research. In M.Brandimonte, G.O.Einstein, & M.McDaniel (Eds.), Prospective memory: Theory and applications (pp. 1–22 ). Mahwah, NJ: Lawrence Erlbaum Associates Inc. Guthke, T., Walther, K., & Thöne-Otto, A.I.T. (2002). Einsatz computerunterstützter Therapieverfahren in der Behandlung von Gedächtnisstörungen. In W.Huber, P.-W. Schönle, P. Weber, & R.Wiechers (Eds.), Computer helfen heilen und leben. Computer in der neurologischen Rehabilitation (S. 123–142 ). Bad Honnef, Germany: Hippocampusverlag Hersh, N.A., & Treadgold, L.G. (1994). NeuroPage: The rehabilitation of memory dysfunction by prosthetic memory and cueing . Neurvpsychological Rehabilitation , 4, 187–197 . Kim, H.J., Burke, D.T., Dowds, M.M., Boone, K.A. R., & Park, G.J. (2000). Electronic memory aids for outpatient brain injury: Follow-up findings. Brain Injury , 14 , 187– 196 . Kinsella, G., Murtagh, D., Landry, A., & Homfray, K. (1996). Everyday memory following traumatic brain injury. Brain Injury , 10 499–507 . Kliegel, M., McDaniel, M.A., & Einstein, G.O. (2000). Plan formation, retention, and execution in prospective memory: A new approach and age-related effects. Memory and Cognition , 28 , 1041–1049 . Nelson, H.E. (1991). National Adult Reading Test (2nd ed). Windsor, UK: The National Foundation for Educational Research. Thöne, A.I.T., & Walther, K. (2001). Neuropsychologische Störungen als Prädiktoren von Selbständigkeit im Alltag. Zeitschrift für Neuropsychologie , 12 , 102–103 . Thöne-Otto, A.I.T., Schulze, H., Irmscher, K., & von Cramon, D.Y. (2001). MEMOS— Interaktive elektronische Gedächtnishilfe für hirngeschädigte Patienten. Deutsches Ärzteblatt, Ausgabe B , 11 , B598–B600 . Van den Broek, M.D., Downes, J., Johnson, Z., Dayus, B., & Hilton, H. (2000). Evaluation of an electronic memory aid in the neuropsychological rehabilitation of prospective memory deficits. Brain Injury , 14 , 455–462 . Wade, T.K., & Troy, J.C. (2001). Mobile phones as a new memory aid: A preliminary investigation using case studies. Brain Injury , 15 , 305–320 . Wechsler, D. (1987). Wechsler Memory Scale-Revised (WMS-R ). New York: The Psychological Corporation. Wilson, B.A. (2002). Memory rehabilitation. Zeitschrift für Neuropsychologie , 13 , 252 .
238
95
Wilson, B.A., Evans, J.J., Emslie, H., & Malinek, V. (1997). Evaluation of NeuroPage: A new memory aid. Journal of Neurology, Neurosurgery and Psychiatry , 63 , 113– 115 . Wright, P., Rogers, N., Hall, C, Wilson, B., Evans, J., Emslie, H., & Bartram, C. (2001). Comparison of pocketcomputer memory aids for people with brain injury. Brain Injury , 15 , 878–800 .
96 HOW TO DESIGN AN ELECTRONIC MEMORY AID
A two-process model of strategic monitoring in event-based prospective memory: Activation/retrieval mode and checking Melissa J.Guynn New Mexico State University, Las Cruces, USA
Theorists have suggested that individuals may remember to execute event-based intended actions by deploying executive or attentional resources to monitor for the markers or target events that indicate that it is appropriate to execute the intended actions (e.g., McDaniel & Einstein, 2000; Shallice & Burgess, 1991), but these strategic monitoring views are not specific about the processes that strategic monitoring entails. A more specific idea is outlined here (see also Guynn, 2001) and an experiment with results consistent with this view is reported. According to this two-process view, strategic monitoring entails maintaining the cognitive system in a prospective memory retrieval mode, which may be mediated by increased activation of the prospective memory representation, plus checking whether the circumstances to execute the intended action are present. In the current experiment, concurrent task impairment on nontarget trials, on which participants were instructed to press a key if they saw a target event (i.e., experimental trials), relative to trials on which participants were not instructed to press a key if they saw a target event (i.e., control trials), provided a footprint of strategic monitoring. An interaction of trial type and whether the experimental and control trials alternated or were blocked revealed greater impairment on experimental trials relative to control trials when the trials were blocked than when the trials alternated. Performance on experimental trials did not vary whether the trials alternated or were blocked, while performance on control trials was significantly worse when the trials alternated than when the trials were blocked. The results are consistent with the two-process view and the idea that participants maintained a retrieval mode/activation and checked on experimental trials, neither maintained a retrieval mode/activation nor checked on blocked control trials, and maintained a retrieval mode/ activation but did not check on alternating control trials. Les théoriciens ont suggéré que les individus peuvent se rappeler d’exécuter des intentions d’actions liées à des événements en
98 GUYNN
déployant des ressources exécutives ou attentionnelles pour surveiller les événements marquants ou ciblés indiquant qu’il est approprié d’exécuter les intentions d’actions (p. ex., McDaniel et Einstein, 2000; Shallice et Burgess, 1991). Mais ces points de vue de surveillance stratégique ne précisent pas le processus sous-jacent a la surveillance stratégique. Dans le present article, une idée plus spécifique est soulevée (voir aussi Guynn, 2001) et une experimentation présentant des résultats supportant ce point de vue est rapportée. Selon la perspective des deux processus, la surveillance stratégique implique, d’une part, un maintien du système cognitif dans un mode de retrait de la mémoire prospective, lequel peut être modifié par l’activation accrue de la representation de la mémoire prospective, et, d’autre part, une verification de la presence des circonstances permettant d’exécuter l’intention d’action. Dans la présente étude, une tâche concurrente comprenant des essais non ciblés dans lesquels les participants ont reçu pour consigne de presser sur un bouton s’ils voyaient un événement-cible (c.-à-d. les essais expérimentaux) et des essais dans lesquels les participants n’ont pas reçu cette consigne (c.-à-d. les essais contrôles) a permis d’identifier la surveillance stratégique. Une interaction du type d’essai et de la façon dont les essais expérimentaux et contrôles alternent ou sont bloqués a révélé une plus grande déficience dans les essais expérimentaux comparativement aux essais contrôles et cette difference est davantage marquée lorsque les essais sont bloqués que lorsqu’ils sont alternés. La performance dans les essais expérimentaux ne varie pas, que les essais soient alternés ou bloqués,
© 2003 International Union of Psychological Science http:// DOI: 10.1080/ www.tandf.co.uk/ 00207590244000205 journals/pp/ 00207594.html
Requests for reprints should be addressed to Melissa J. Guynn, Department of Psychology, MSC 3452, New Mexico State University, PO Box 30001, Las Cruces, NM 88003–8001, USA (Email:
[email protected]). Appreciation is expressed to Roger Chadwick and Jaime Ortiz for experimental assistance, and to Gil Einstein, Fred Mast, Mark McDaniel, and Dom Simon for helpful comments on a prior version of the manuscript.
STRATEGIC MONITORING IN PROSPECTIVE MEMORY
tandis que la performance dans les essais contrôles est significativement moins bonne quand les essais sont alternés que lorsqu’ils sont bloqués. Les résultats sont en accord avec la perspective des deux processus et l’idée que les participants maintiennent un mode de retrait/activation et vérifient les essais expérimentaux; ne maintiennent pas un mode de retrait/activation et ne vérifient pas les essais contrôles bloqués; et maintiennent un mode de retrait/activation sansvérifier les essais contrôles alternés. Los individuos pueden recordar la ejecución acciones basadas en eventos, de acuerdo con algunos teóricos, por medio del desarrollo de recursos ejecutivos o de la atención para verificar los eventos meta que indiquen que es apropiado ejecutar las acciones intentadas (por ejemplo, McDaniel & Einstein, 2000; Shallice & Burgess, 1991), pero estos puntos de vista de la verificación estratégica no especifican el proceso que la verificación estratégica conlleva. Aquí se delinea una idea más específica (véase también Guynn, 2001) y se informa sobre un experimento, cuyos resultados son consistentes con esta perspectiva. De acuerdo con esta vision de dos procesos, la verificación estratégica implica que el sistema cognitivo se mantiene en estado de preparación para la recuperación de la memoria prospectiva, mediado tal vez por la activación aumentada de la representación de la memoria prospectiva, además de que verifica si están presentes las circunstancias para ejecutar la acción intentada. En el presente experimento, los ensayos con una tarea concurrente en los que se instruía a los participantes a oprimir una tecla si veían el evento meta (ensayos experimentales) en relación con ensayos cuyos participantes no recibían la instrucción de oprimir la tecla si veían el evento meta (ensayos de control) proporcionaron el sello de identidad de la verificación estratégica. La interacción entre el tipo de ensayo y si los ensayos experimentales y de control alternaban o se bloqueaban revelaron mayor deficiencia en los ensayos experimentales que en los de control cuando los ensayos se bloqueaban que cuando alternaban. El desempeño en los ensayos experimentales no varió entre ensayos que alternaban o se bloqueaban, pero en los ensayos control el desempeño fue significativamente peor cuando los ensayos alternaban que cuando se bloqueaban. Los resultados son consistentes con la perspectiva de dos procesos y con la idea de que los participantes mantenían una preparación para la recuperación/activación y verificaban en los ensayos experimentales, no mantenían una preparación para la recuperación/activación ni verificaban en los ensayos de control, y mantenían una preparación para la recuperación/activación pero no verificaban en los ensayos de control que alternaban.
99
100 GUYNN
Event-based prospective memory refers to memory to execute an intended action upon the occurrence of a designated target event that indicates that it is appropriate to execute the intended action (Einstein & McDaniel, 1990; Meacham & Leiman, 1982). An example in the life of this author is remembering to write down interesting cases of prospective memory successes and failures as they occur in everyday life, so they can be given as examples in a manuscript at a later date (she often fails to remember this intention). A question that has motivated recent prospective memory research is how an individual remembers to execute an intended action upon the occurrence of a designated target event, given that there is no explicit request to remember the intention. On the one hand, remembering under these circumstances might depend on relatively strategic processes, given that not only must one remember the intention, but also that one must remember the intention at the appropriate point (i.e., one must remember to remember; Craik, 1986; McDaniel & Einstein, 2000). On the other hand, remembering under these circumstances might depend on relatively automatic processes, with the occurrence of the designated target event initiating automatic or reflexive retrieval of the intended action (McDaniel & Einstein, 2000). Several theoretical perspectives that reflect these possibilities have been offered in recent years, ranging from viewing prospective memory as relatively automatic (e.g., Automatic Associative Activation view, Noticing +Search view; Einstein & McDaniel, 1996; Guynn, McDaniel, & Einstein, 2001; McDaniel, 1995; McDaniel & Einstein, 2000; McDaniel, Guynn, Einstein, & Breneiser, 2003) to viewing prospective memory as relatively strategic (e.g., Strategic Monitoring view; Burgess & Shallice, 1997; Ellis, 1996; McDaniel & Einstein, 2000; Shallice & Burgess, 1991; Smith & Bayen, 2002a). A multiprocess view has also been proposed, suggesting that the extent to which prospective memory retrieval is mediated by automatic versus strategic processes may vary and depend on characteristics of the prospective memory task, the cover task in which the prospective memory task is embedded, and the individual doing the prospective remembering (McDaniel & Einstein, 2000; see also Einstein, McDaniel, Shank, & Mayfield, 2002). STRATEGIC MONITORING Although several perspectives suggest that prospective memory may be mediated by strategic monitoring, the views are not specific about the processes that strategic monitoring entails (cf. Smith & Bayen, 2002a). Accordingly, the purpose of this article is to introduce an idea about these speciflc processes (see also Guynn, 2001) and to report an experiment that provides evidence for this view as well as for the more general strategic monitoring view. The strategic monitoring view of prospective memory (McDaniel & Einstein, 2000; see also Smith & Bayen, 2002a) is based on Shallice and Burgess’ (1991; Burgess & Shallice, 1997) proposal that an executive attentional system or
STRATEGIC MONITORING IN PROSPECTIVE MEMORY
101
supervisory attentional system (SAS) is used to control behaviour when welllearned triggering procedures are not sufficient to guarantee an intended outcome. Thus a SAS is used to control behaviour when there is a prospective memory task, because such a task is unlikely to correspond to a well-learned action plan that can be retrieved automatically. The SAS is used to encode the markers or target events that indicates that it is appropriate to execute the intended action, to monitor for the markers or target events, and to interrupt the ongoing activity when a marker or target event occurs so that the intended action can be executed. McDaniel and Einstein (2000) acknowledged that the process of monitoring has not been well detailed but suggested that executive processing resources could be used to continuously monitor the environment for the target events, or to periodically bring the intended action to mind, which could maintain the activation of the association between the target event and the intended action. Smith and Bayen (2002a) suggested another possibility—that preparatory attentional processes (and retrospective memory processes) mediate prospective memory. The view of strategic monitoring introduced here is consistent with the general strategic monitoring view (in that it assumes that monitoring depends on limited processing capacity) but draws inspiration from Tulving’s (1983) proposal that retrieval depends on two things—the cognitive system being in a retrieval mode, and there being a retrieval cue (external or internal) to initiate ecphoric retrieval (i.e., the interaction of a retrieval cue with a stored memory trace). RETRIEVAL MODE A retrieval mode is a cognitive or neurocognitive task set to treat stimuli as cues to retrieve stored episodes, and it is separate from the actual ecphory or retrieval of the stored episodes (Duzel, 2000; Nyberg, 1999; Tulving, 1983, 1998). Behavioural evidence has been provided by Craik, Naveh-Benjamin, Anderson, and colleagues (N. D.Anderson, Craik, & Naveh-Benjamin, 1998; Craik, Govoni, Naveh-Benjamin, & Anderson, 1996; Craik, Naveh-Benjamin, Ishaik, & Anderson, 2000; Naveh-Benjamin, Craik, Gavrilescu, & Anderson, 2000) in studies on the effects of dividing attention at encoding and retrieval in free recall, cued recall, and recognition. Compared to dividing attention at encoding, dividing attention at retrieval produces little to no reduction in the number of items retrieved, suggesting that retrieval/ ecphory is relatively automatic. Further, emphasizing at retrieval that one of the tasks (either the memory retrieval task or the divided attention task) is more important does not affect the number of items retrieved. Finally, compared to dividing attention at encoding, dividing attention at retrieval produces greater impairment on the divided attention task, suggesting that being in a retrieval mode is not relatively automatic and instead depends on limited processing capacity.
102 GUYNN
Neuropsychological evidence for a retrieval mode has also been provided. Positron emission tomography (PET) studies have revealed activity in the right prefrontal cortex when participants attempt to retrieve studied information, regardless of whether the retrieval attempt succeeds or fails (e.g., McIntosh, Nyberg, Bookstein, & Tulving, 1997; Nyberg et al, 1995). An event-related potential (ERP) study has revealed a sustained positivity on trials on which a cue signals an upcoming episodic memory (recognition) test trial as compared to an upcoming semantic memory (animacy judgment) test trial, regardless of whether the retrieval attempt succeeds or fails (Morcom & Rugg, 2002). The positivity begins about 500 ms after the cue signals the upcoming recognition trial, but not until the second successive trial, suggesting that adopting a task set takes time (see also Duzel et al., 1999, cited in Duzel, 2000). As revealed by these studies, both behavioural and neuropsychological evidence suggest that the distinction between a retrieval mode and the actual ecphory or retrieval of stored episodes may be a useful characterization of retrospective memory retrieval (i.e., retrieval of information encountered in the past in response to an explicit request to remember). The view of strategic monitoring introduced here suggests that the distinction between a retrieval mode and the actual ecphory or retrieval of stored episodes may also be a useful characterization of strategic monitoring in event-based prospective memory. In much the same way that being in a retrieval mode may be a prerequisite for retrieving stored retrospective episodes (Tulving, 1983), being in a retrieval mode may also be a prerequisite for retrieving stored prospective episodes. A prospective retrieval mode would be a cognitive or neurocognitive task set to treat stimuli (external or internal) as cues to retrieve intentions. A prospective retrieval mode may differ from a retrospective retrieval mode (e.g., a retrieval attempt; Tulving, 1998), however, and may be best characterized as a preparedness or a readiness to treat stimuli as retrieval cues. What it entails to be in a retrieval mode in a mechanistic sense has been difficult to specify (e.g., “We know next to nothing about retrieval mode,” Tulving, 1983, p. 169; “The functions and characteristics of the retrieval mode remain obscure,” Craik et al., 1996, p. 177), but with regard to prospective memory, one candidate is the increased level of subthreshhold activation of the prospective memory representation in memory (Goschke & Kuhl, 1993, 1996; Marsh, Hicks, & Bink, 1998; Marsh, Hicks, & Bryan, 1999; Marsh, Hicks, & Watson, 2002) 1 . Similarities between characteristics of a retrieval mode and the increased level of activation of prospective memory representations provide some evidence for a relationship between the two. Evidence and theorists have suggested that retrieval modes and increased levels of activation persist until stored items are retrieved or until the retrieval goal changes, can be effortful to establish and maintain, can be effortful to switch in and out of, may depend on limited processing capacity, and may depend on the frontal cortex (Duzel, 2000; Goschke & Kuhl, 1993, 1996; Marsh et al., 1998, 1999, 2002; McIntosh et al.,
STRATEGIC MONITORING IN PROSPECTIVE MEMORY
103
1997; Morcom & Rugg, 2002; Nyberg, 1999; Nyberg et al., 1995; Tulving, 1983, 1998). ACTIVATION Goschke and Kuhl (1993, 1996) were the first to suggest that prospective memory representations are characterized by stronger and longer-lasting levels of subthreshhold activation than are comparable retrospective memory representations. Evidence came from experiments in which participants were asked to learn scripts of actions (e.g., distribute the cutlery in the setting a dinner table script) that they expected later either to perform or to recall. Words from both types of scripts were then intermixed with nonpresented lures for a speeded recognition test. Words from the to-be-performed scripts were recognized faster than words from the to-be-recalled scripts. Marsh, Hicks, and colleagues (Marsh et al., 1998, 1999, 2002; see also Cohen, Lindsay, & Dixon, 2002, using a Stroop task) obtained a similar intention superiority effect using a lexical decision task and extended the results with evidence for increased activation when participants expected to perform the actions in the future (i.e., for uncompleted and partially completed intentions) but not when the actions had been completed or cancelled. Goschke and Kuhl (1993, 1996; see also Marsh et al., 1998, 1999) interpreted the intention superiority effect within the framework of J.R. Anderson’s (1983) adaptive control of thought (ACT*) model of cognition, which proposes that goal nodes in working memory sustain increased levels of activation without rehearsal. Goschke and Kuhl (1993, 1996) suggested that the representations of intentions (i.e., prospective memory representations) might also be goal nodes that sustain increased levels of activation without rehearsal. More recently, Lebiere and Lee (2001) modelled the intention superiority effect using the ACTR cognitive architecture (a successor of ACT*; J.R.Anderson & Lebiere, 1998). In this model, processing depends on the goal of the cognitive system (i.e., the goal to perform a particular task). Information about the current stimulus is included in the goal, and information about the current context or task to be performed is included as a source of activation in the goal. When the task or context is changed, the level of activation of the stimuli that are most likely to be encountered in the old task or context is decreased, and the level of activation of the stimuli that are most likely to be encountered in the new task or context is increased. An important assumption in this computational account of the intention superiority effect (particularly to the current application to strategic monitoring) is that maintaining a task or context in the goal demands limited attentional
1
Alternatively, an increased level of subthreshhold activation may be the mechanism of strategic monitoring, and it may or may not instantiate a prospective memory retrieval mode.
104 GUYNN
processing capacity (see also, e.g., Lovett, Reder, & Lebiere, 1999). Prospective memory researchers have attributed the intention superiority effect to an inherent increased level of activation for prospective versus retrospective representations — prospective representations sustain activation without rehearsal (Goschke & Kuhl, 1993, 1996; Marsh et al., 1998,1999,2002). If no strategic process (e.g., rehearsal) plays a role in the intention superiority effect, then increased levels of activation seem improbable as a mechanism of strategic monitoring, but if a strategic process (i.e., establishing and maintaining a task or context in the goal) plays a role, then increased levels of activation seem more plausible as a mechanism of strategic monitoring. CHECKING The second component of this view of strategic monitoring inspired by Tulving (1983) is checking and is based on the idea that a retrieval cue (external or internal) initiates ecphoric retrieval. In event-based prospective memory, the retrieval cue would most likely be the target event that indicates that it is appropriate to execute the intended action. This component of strategic monitoring would involve checking the environment for the target events, which would involve attention to the stimuli where the retrieval cue would be expected to occur (e.g., in the context of the cover task in which the prospective memory task is embedded) and evaluation of whether or not a stimulus in that context is a retrieval cue for an intended action. CURRENT EXPERIMENT An experiment was conducted to evaluate this two-process view of strategic monitoring using a method that was developed to reveal footprints of strategic monitoring in event-based prospective memory (Guynn, 2001). The method was inspired by work by Craik, Naveh-Benjamin, Anderson, and colleagues (N.D. Anderson et al., 1998; Craik et al., 1996,2000; Naveh-Benjamin et al., 2000) and involves several modifications to the method of typical prospective memory experiments. In typical experiments, a prospective memory task is embedded in a cover task, such that prospective memory target events represent some (usually small) proportion of the cover task stimuli. In the current experiment, the cover task was a short-term memory task based on a task developed by Einstein and McDaniel (1990; see also Einstein, Holland, McDaniel, & Guynn, 1992, and McDaniel & Einstein, 1993), and the prospective memory task was to press the Enter key if a word that was a fruit appeared in a short-term memory trial. The goal was to simulate aspects of real-world prospective remembering in which an intention must be retrieved while an individual is engaged in another activity. For example, seldom does one have the luxury of doing nothing but awaiting the appearance of a colleague so a message can be relayed. It is more often the case that one is reading a journal article, hurrying to a lab meeting, or talking with a
STRATEGIC MONITORING IN PROSPECTIVE MEMORY
105
student when the colleague appears, and thus the intention to relay the message must be retrieved while one’s attention is directed toward the other activity. Several modifications to the method of typical prospective memory experiments were introduced to reveal footprints of strategic monitoring (Guynn, 2001). First, a reaction time task was performed at the same time as the short-term memory task. The reaction time task involved asterisks being presented in one of four positions on the computer screen, and participants pressing one of four computer keys as quickly and accurately as possible to indicate the position of the asterisks. This type of reaction time task demands strategic processing resources (e.g., Craik et al., 1996); if participants adopt a strategy of monitoring for the target events, strategic monitoring will take processing resources away from performing the reaction time task and performance on the reaction time task will be impaired. Second, participants were not always instructed to execute an intended action if a designated target event appeared (i.e., participants were not always given a prospective memory task). In particular, during control trials, participants were instructed just to perform the short-term memory task and the reaction time task, but during experimental trials, participants were instructed to perform all three tasks together—the short-term memory task, the reaction time task, and the prospective memory task. Finally, a fruit appeared on only a small proportion of the experimental trials (e.g., 4 out of 48 trials) and thus participants did not press the Enter key on every experimental trial. The trials on which a fruit appeared and the trials on which a participant pressed the Enter key were omitted from the analyses of reaction time task performance, because the interest was in whether participants adopted a strategy of monitoring for the target events, uncontaminated by prospective memory retrieval and/or a motor response. In prior work using this task (e.g., Guynn, 2001, Experiment 2), the experimental and control trials were blocked by trial type, such that participants received either 48 experimental trials followed by 48 control trials, or 48 control trials followed by 48 experimental trials. Reaction time task performance was impaired on experimental trials relative to control trials, and the impairment was interpreted as a footprint of strategic monitoring and as evidence for the general strategic monitoring view of prospective memory (McDaniel & Einstein, 2000; see also Smith & Bayen, 2002a). The impairment is also consistent with the twoprocess view of strategic monitoring introduced here, as well as with two alternate one-process views—one view in which maintaining the cognitive system in a prospective memory retrieval mode (or maintaining an increased level of activation of the prospective memory representation) is the only component of strategic monitoring, and one view in which checking the environment for the target events is the only component of strategic monitoring. A modification of Guynn’s (2001) experimental procedure allows different predictions to be made for the two-process view and the alternate one-process views. Specifically, the three views make different predictions about reaction time
106 GUYNN
task performance on experimental trials versus control trials as a function of whether the experimental and control trials alternate or are blocked by trial type. PREDICTIONS According to the two-process view, participants would be expected to maintain a retrieval mode (or an increased level of activation) and to check for the target events on experimental trials. In contrast, participants would not be expected to maintain a retrieval mode (or an increased level of activation) nor to check for the target events on blocked control trials. Participants would be expected to maintain a retrieval mode (or an increased level of activation) on alternating control trials, however, based on prior research that suggests that a retrieval mode (or an increased level of activation) persists until the retrieval task ends, but participants would not be expected to check for the target events. If both processes demand limited processing capacity, then reaction time task performance should be worst on experimental trials, best on blocked control trials, and intermediate on alternating control trials. Accordingly, performance on experimental trials relative to control trials should depend on whether the experimental and control trials alternate or are blocked—an interaction of the factors would be expected. The monitoring effect (costs on experimental trials relative to control trials) was expected to be larger when the trial types were blocked than when the trial types alternated. Moreover, the interaction was expected to arise because of differential performance on alternating versus blocked control trials rather than alternating versus blocked experimental trials. A second possibility is that there is no process of checking. It may be that just a process of maintaining a retrieval mode (or an increased level of activation) mediates strategic monitoring. Participants would be expected to maintain a retrieval mode (or an increased level of activation) on experimental trials and on alternating control trials (i.e., until the retrieval task ends). In contrast, participants would not be expected to maintain a retrieval mode (or an increased level of activation) on blocked control trials. Reaction time task performance should be worst on experimental trials and alternating control trials and best on blocked control trials. Accordingly, performance on experimental trials relative to control trials should depend on whether the experimental and control trials alternate or are blocked—an interaction of the factors would be expected. But unlike the two-process monitoring view prediction, the monitoring effect on the blocked trials was expected to disappear altogether when the trial types alternated. Moreover, the interaction was expected to arise because of differential performance on alternating versus blocked control trials rather than alternating versus blocked experimental trials. A third possibility is that there is no process of maintaining a retrieval mode (or an increased level of activation). It may be that just a process of checking mediates strategic monitoring. Participants would be expected to check for the target events on experimental trials but not on control trials, and reaction time
STRATEGIC MONITORING IN PROSPECTIVE MEMORY
107
task performance should be worst on experimental trials and best on control trials. Accordingly, performance on experimental trials relative to control trials should not depend on whether the experimental and control trials alternate or are blocked—no interaction of the factors would be expected. These three possibilities (strategic monitoring involves both activation/ retrieval mode and checking, just activation/retrieval mode, or just checking) were evaluated in an experiment in which experimental and control trials alternated and were blocked by trial type. The interest was in the extent of monitoring costs or footprints (i.e., experimental versus control trial differences on the reaction time task) on blocked versus alternating trials. METHOD Design The design was a 2×2 repeated measures factorial varying trial type (control, experimental) and trial presentation (blocked, alternating). For the control trials, participants performed the short-term memory and reaction time tasks, and for the experimental trials, participants performed the short-term memory, reaction time, and prospective memory tasks. For the blocked trials, participants performed 24 consecutive control trials and 24 consecutive experimental trials, and for the alternating trials, participants performed 48 alternating control and experimental trials. Four different orders of the treatment conditions were used. Either the blocked trials preceded the alternating trials or vice versa, and the first trial was either a control trial or an experimental trial. Participants Participants were 64 undergraduates enrolled in psychology courses at New Mexico State University who participated either in partial fulfilment of a course requirement or for extra credit. There were 16 participants in each counterbalancing condition. Materials Stimuli for the short-term memory trials were 540 one-syllable and twosyllable common nouns selected from Clusters 6, 7, and 8 of the Toglia and Battig (1978) norms. For the test trials, 480 words were arranged into 96 shortterm memory trials (16 blocks of 6 trials). Five words were selected randomly without replacement for each trial, with the restriction that there were no more than three 2-syllable words in a trial. A prospective memory target event (apple, peach, banana, plum, cherry, grape, apricot, pear) appeared in 8 of the 48 experimental trials— one target event in each of the four blocks of blocked
108 GUYNN
experimental trials, and one target event in four of the eight blocks of alternating trials. For the practice trials, 60 words were arranged into two blocks of six trials. One block was used for practising the short-term memory task, and the other block was used for practising the short-term memory and reaction time tasks together, and for practising the short-term memory, reaction time, and prospective memory tasks together. Four practice target events (rake, truck, nose, soap) appeared in the block used for practising the prospective memory task. Procedure The experiment lasted about 1 hour. Participants were informed that the experiment required performing several different tasks together and that they would practise the tasks before beginning the actual test trials. Participants were first given instructions and practice for the reaction time task. Asterisks would appear in one of four positions displayed horizontally on the computer screen. If asterisks appeared in the first position (on the far left) participants had to press the z key, in the second position they had to press the x key, in the third position they had to press the n key, and in the fourth position (on the far right) they had to press the m key. As soon as the correct key was pressed, the asterisks would move to one of the other three locations, chosen randomly. Participants had to press the keys as quickly and accurately as possible, as long as the asterisks were presented, using the first two fingers of each hand. For each of six practice trials, asterisks were presented continuously in this manner (upon a correct key press) for 9.5 s, which was the duration of a short-term memory trial. Participants were next given instructions and practice for the short-term memory task. Five words would appear in a row in the centre of the computer screen for 5 s. Participants should study the words while they appeared, and say as many as possible out loud and in order when they disappeared. For each of six practice trials, participants were given 5 s to study the words and 4.5 s to recall the words. Participants were next instructed that they would practise the reaction time task and the short-term memory task together. In particular, they should press the keys corresponding to the asterisks while studying the words and while recalling the words. Participants were also instructed that before each trial, a 2 would appear at the top of the screen as a reminder to do the two tasks together. Participants then received two identical blocks of six practice trials. The 2 appeared on the screen for 1.5 s before and during each 9.5 s trial. Participants were next instructed that a third task would be performed with the reaction time task and the short-term memory task. Participants were instructed to press the Enter key if they saw rake or truck as a word in a short-term memory trial. Participants were also instructed that before each trial, a 3 would appear at the top of the screen as a reminder to do the three tasks together. Participants then received six practice trials. The 3 appeared on the screen for 1.5 s before
STRATEGIC MONITORING IN PROSPECTIVE MEMORY
109
and during each 9.5 s trial. Participants were next instructed to press the Enter key if they saw nose or soap as a word in a short-term memory trial and they received the six practice trials again. (The same block of six practice trials was used four times so that familiarity with the words would enable participants to become proficient at performing the tasks before beginning the actual test trials.) Participants were next informed that the test trials would be organized into blocks of six trials, with 10-s rest periods between the blocks. Participants would perform the reaction time and short-term memory tasks together on control trials, and the reaction time, short-term memory, and prospective memory tasks together on experimental trials. For the prospective memory task, participants were instructed to press the Enter key if they ever saw a word that was a fruit. This instruction was given once before the alternating trials and once before the blocked experimental trials. Participants received 96 short-term memory test trials, arranged into 16 blocks of 6 trials. The first 48 trials consisted of either 24 control trials followed by 24 experimental trials, 24 experimental trials followed by 24 control trials, 48 alternating control and experimental trials, starting with a control trial, or 48 alternating control and experimental trials, starting with an experimental trial. The last 48 trials consisted of alternating trials for participants who had already received blocked trials, and blocked trials for participants who had already received alternating trials. The last 48 trials began with the same trial type as the first 48 trials. Participants were informed when the trial types would alternate and when the trial types would be blocked. At the beginning of the first 48 trials, and at the beginning of the last 48 trials, participants were reminded that either a 2 or a 3 would appear before each trial to help them remember whether they should do all three tasks or just the first two tasks they had practised. RESULTS All reported effects were significant at a .05 level unless otherwise indicated. Reaction time task The number of correct asterisk key presses and the reaction times for correct asterisk key presses during the study phase of the short-term memory task were used as converging measures of reaction time task performance (Guynn, 2001). For the accuracy measure, higher values reflect better performance, and for the latency measure, lower values reflect better performance. Although the measures are referred to as accuracy and latency, each measure reflects information about both the accuracy and the latency of the key presses. The accuracy measure is higher if participants are more accurate, and if they are quicker, because quicker key presses leave more time for additional key presses. The latency measure is lower if participants are quicker, and if they are more accurate, because only correct key presses are recorded and cause the asterisks to move, so an incorrect
110 GUYNN
key press is recorded as a longer reaction time on the next correct key press. It is not necessary to get significant effects on both measures, as long as the effects on the two measures reflect the same pattern. Following N.D.Anderson et al. (1998), key presses with a latency of less than 100 ms were excluded from the analyses. The mean number of correct asterisk key presses per treatment condition is presented in Figure 1. The results were analysed with a 2×2 repeated measures analysis of variance, with trial type (control, experimental) and trial presentation (blocked, alternating) as factors. There were significant main effects of trial type, control >experimental, F(1, 63)= 97.55, MSE=0.20, and trial presentation, blocked > alternating, F(1, 63)=13.34, MSE=0.23, in the context of a significant interaction, F(1, 63)=6.32, MSE= 0.11. The interaction reflected a greater difference between control and experimental trials when the trial types were blocked than when the trial types alternated, and a greater difference between blocked and alternating control trials than between blocked and alternating experimental trials. In fact, there was no significant difference between blocked and alternating experimental trials, F(1, 63)=0.88, MSE= 0.11, but there was a significant difference between blocked and alternating control trials, F(1, 63)=7. 45, MSE=0.11. The mean reaction time for correct asterisk key presses per treatment condition is presented in Figure 2. The results were analysed with a 2×2 repeated measures analysis of variance, with trial type (control, experimental) and trial presentation (blocked, alternating) as factors. There were significant main effects of trial type, control < experimental, F(l, 63)= 54.68, MSE=183 89.10, and trial presentation, blocked < alternating, F(1, 63)=9.16, MSE=19468.89, and1 although the interaction was not significant, F(1, 63)= 0.21, MSE=6435.75, the pattern was similar to that in Figure 1. There was a greater difference between control and experimental trials when the trial types were blocked than when the trial types alternated, and a greater difference between blocked and alternating control trials than between blocked and alternating experimental trials. In fact, there was a significant difference between blocked and alternating control trials, F(1, 63)=4. 09, MSE=6435.75, but there was no significant difference between blocked and alternating experimental trials, F(1, 63)=2.89, MSE=6435.75. Prospective memory There was not a significant difference in prospective memory accuracy (average proportion of correct prospective memory responses) as a function of whether the experimental trials were blocked (.84) or alternated (.79) with the control trials, F(1, 63)—1.88, MSE=0.04. There also was not a significant difference in prospective memory latency (average reaction time for correct prospective memory responses) as a function of whether the experimental trials were blocked
STRATEGIC MONITORING IN PROSPECTIVE MEMORY
111
Figure 1. Mean reaction time task accuracy per treatment condition.
Figure 2. Mean reaction time task latency per treatment condition.
(3074.96 ms) or alternated (2992.36 ms) with the control trials, F(1, 62)=0.35, MSE=441221.78. DISCUSSION The results suggest that two processes mediate strategic monitoring in eventbased prospective memory—one process that can be turned on and off fairly easily, which was operating during experimental trials but not control trials, and one process that cannot be turned on and off as easily, which was operating during experimental trials and alternating control trials but not blocked control trials. The results were thereby consistent with the two-process view of strategic monitoring introduced here, with one process corresponding to checking the environment for the target events and one process corresponding to maintaining the cognitive system in a prospective memory retrieval mode (or maintaining an increased level of activation of the prospective memory representation).
112 GUYNN
According to the two-process view, participants were expected to maintain the cognitive system in a prospective memory retrieval mode (or maintain an increased level of activation of the prospective memory representation) and to check for the target events on experimental trials. Accordingly, performance on the reaction time task was worst on these trials. Participants should not have expected target events to appear during blocked control trials and thus were not expected either to maintain the cognitive system in a prospective memory retrieval mode (or maintain an increased level of activation of the prospective memory representation) or to check for the target events. Accordingly, performance on the reaction time task was best on these trials. Participants were expected to maintain the cognitive system in a prospective memory retrieval mode (or maintain an increased level of activation of the prospective memory representation) on alternating control trials, based on prior research that suggests that a retrieval mode (or an increased level of activation) persists until the retrieval task ends, but participants should not have expected target events to appear and thus were not expected to check for them. Accordingly, performance on the reaction time task was intermediate on these trials. The intermediate level of reaction time task performance on alternating control trials could have resulted because some participants monitored on alternating control trials (i.e., just as on experimental trials) and some participants did not monitor on alternating control trials (i.e., just as on blocked control trials). If participants adopted these different strategies, then the frequency distributions of the reaction time task accuracies and latencies would have been bimodal, with participants who monitored having a lower average accuracy and a higher average latency than participants who did not monitor. Neither frequency distribution appeared to be bimodal, however, consistent with the expectation that participants maintained the cognitive system in a prospective memory retrieval mode (or maintained an increased level of activation of the prospective memory representation) but did not check for the target events on alternating control trials. A significant difference in prospective memory accuracy or latency was not expected (and was not obtained) as a function of whether the experimental and control trials alternated or were blocked. The two-process view of strategic monitoring, as well as the alternate one-process views, predict that strategic monitoring should be equivalent on the alternating and blocked experimental trials. The views make different predictions only with regard to the processes that should be involved on the alternating versus blocked control trials, where prospective memory cannot be measured (i.e., because there is no prospective memory task on control trials). CHECKING The current experiment was not designed to test alternate conceptions of the checking process, but several possible conceptions are outlined here. Depending
STRATEGIC MONITORING IN PROSPECTIVE MEMORY
113
on the characteristics of the prospective memory task, the cover task in which it is embedded, and the individual doing the strategic monitoring, the process of directing attention to the stimuli could be controlled by the environment and be relatively automatic or reflexive on the part of the individual (e.g., an exogenous orienting response; Lauwereyns, 1998), or instead it could be controlled by the individual or a SAS or other type of executive attentional system (e.g., an endogenous orienting response; Lauwereyns, 1998). Alternatively, some aspect of automatic memory retrieval, such as automatic associative retrieval of the intended action, which results from conscious processing of the target event with which it was associated at encoding (as postulated by the Automatic Associative Activation view of prospective memory; Einstein & McDaniel, 1996), or a feeling of familiarity or significance for the target event that causes the target event to be noticed (as postulated by the Noticing+Search view of prospective memory; Einstein & McDaniel, 1996), could direct attention to the stimuli. The process of evaluating whether a stimulus is a retrieval cue for an intended action could also be relatively automatic or strategic. For example, automatic or strategic processes thought to underlie recognition memory (e.g., Mandler, 1980; Sternberg, 1969) could play a role. Alternatively, some aspect of memory retrieval as postulated by the Automatic Associative Activation view or the Noticing+Search view (Einstein & McDaniel, 1996) could be the process by which the evaluation is made. Finally, to the extent that the ACT architecture (J.R.Anderson, 1983) can account for strategic monitoring, the firing of production rules could explain the process of checking. Although these various possible conceptions of the checking process allow that checking may be automatic or strategic, the reaction time task costs on experimental trials relative to alternating control trials in the current experiment suggest that checking was relatively strategic, at least in the current experiment. ALTERNATE ONE-PROCESS INTERPRETATIONS Although the two-process monitoring view provides a compelling explanation of the current results, the results are open to alternate one-process interpretations. One possibility is that the costs on the experimental trials and the alternating control trials relative to the blocked control trials reflect only a process of maintaining a retrieval mode (or activation): Participants maintained a retrieval mode (or activation) to a greater extent on experimental trials than alternating control trials, and to a greater extent on alternating control trials than blocked control trials, where they were not expected to maintain a retrieval mode at all (or at least to a much lesser extent). This seems unlikely, because research has suggested that a retrieval mode or an increased level of activation persists as long as the goal to perform the retrieval task exists, and there was no reason to expect that a retrieval mode or activation on alternating control trials would be maintained to a lesser extent than on experimental trials.
114 GUYNN
Another possibility is that the costs on the experimental trials and the alternating control trials relative to the blocked control trials reflect only a process of checking: Participants checked for the target events to a greater extent on experimental trials than on alternating control trials, and to a greater extent on alternating control trials than on blocked control trials, where they were not expected to check at all. This seems unlikely, because checking was costly, and a signal was given to indicate that no target event would appear on the control trials and that checking was therefore unnecessary, so there was no reason to expect checking on any control trials. CONCLUSIONS The current project represents an initial step in trying to understand the processes that mediate strategic monitoring in event-based prospective memory. To the extent that the view of strategic monitoring introduced here captures important processes, future work is likely to focus profitably on issues such as: the relationship between retrieval mode and activation; the nature of checking; the relationship of retrieval mode/activation and checking, including whether they are independent or whether one is a prerequisite for the other; the relationship of retrieval mode/activation and checking to prospective memory; the factors that influence the extent to which retrieval mode/activation and checking are recruited to support prospective memory; and the factors that influence the extent to which retrieval mode/ activation and checking depend on limited processing capacity. In turn, the answers to these and other questions should help to elaborate and refine the view introduced here. Although this view is consistent with the results of the current experiment, no doubt other views may be consistent with the results as well. The view introduced here was offered as a possible more specific instantiation of the more general strategic monitoring view (McDaniel & Einstein, 2000; Shallice & Burgess, 1991; Smith & Bayen, 2002a) that has motivated a lot of recent work on prospective memory (e.g., Einstein et al, 2002; Guynn, 2001; Kliegel, Martin, McDaniel, & Einstein, 2001; Smith, 2000; Smith & Bayen, 2002b), with the goal of stimulating new work that may result in a more complete understanding of prospective memory in general and strategic monitoring in event-based prospective memory in particular. REFERENCES Anderson, J.R. (1983). The architecture of cognition . Cambridge, MA: Harvard University Press. Anderson, J.R., & Lebiere, C. (1998). The atomic components of thought . Mahwah, NJ: Lawrence Erlbaum Associates Inc.
STRATEGIC MONITORING IN PROSPECTIVE MEMORY
115
Anderson, N.D., Craik, F.I.M., & Naveh-Benjamin, M. (1998). The attentional demands of encoding and retrieval in younger and older adults: 1. Evidence from divided attention costs. Psychology and Aging , 13 , 405–423 . Burgess, P.W., & Shallice, T. (1997). The relationship between prospective and retrospective memory: Neuropsychological evidence. In M.A. Conway (Ed.), Cognitive models of memory ( pp. 247–272 ). Cambridge, MA: MIT Press. Cohen, A.-L., Lindsay, D.S., & Dixon, R.A. (2002, November). The intention superiority effect: Instructions to complete or forget an intention modulates Stroop performance . Poster presented at the 43rd annual meeting of the Psychonomic Society, Kansas City, MO. Craik, F.I.M. (1986). A functional account of age differences in memory. In F.Klix & H.Hagendorf (Eds.), Human memory and cognitive capabilities: Mechanisms and performances ( pp. 409–422 ). Amsterdam: Elsevier. Craik, F.I.M., Govoni, R., Naveh-Benjamin, M., & Anderson, N.D. (1996). The effects of divided attention on encoding and retrieval processes in human memory. Journal of Experimental Psychology: General , 125 , 159–180 . Craik, F.I.M., Naveh-Benjamin, M., Ishaik, G., & Anderson, N.D. (2000). Divided attention during encoding and retrieval: Differential control effects? Journal of Experimental Psychology: Learning, Memory, and Cognition , 26 , 1744–1749 . Duzel, E. (2000). When, where, what: The electromagnetic contribution to the WWW of brain activity during recognition. Acta Psychologica , 105 , 195–210 . Einstein, G.O., Holland, L.J., McDaniel, M.A., & Guynn, M.J. (1992). Age-related deficits in prospective memory: The influence of task complexity. Psychology and Aging , 7, 471–478 . Einstein, G.O., & McDaniel, M.A. (1990). Normal aging and prospective memory. Journal of Experimental Psychology: Learning, Memory, and Cognition, 16 , 717– 726 . Einstein, G.O., & McDaniel, M.A. (1996). Retrieval processes in prospective memory: Theoretical approaches and some new empirical findings. In M.Brandimonte, G. O.Einstein, & M.A.McDaniel (Eds.), Prospective memory: Theory and applications ( pp.115–142 ). Mahwah, NJ: Lawrence Erlbaum Associates Inc. Einstein, G., McDaniel, M., Shank, H., & Mayfield, S. (2002, November). The costs of performing a prospective memory task on cover activities: Support for the multiprocess view . Poster presented at the 43rd annual meeting of the Psychonomic Society, Kansas City, MO. Ellis, J. (1996). Prospective memory or the realization of delayed intentions: A conceptual framework for research. In M.Brandimonte, G.O.Einstein, & M.A.McDaniel (Eds.), Prospective memory: Theory and applications ( pp. 1–22 ). Mahwah, NJ: Lawrence Erlbaum Associates Inc. Goschke, T., & Kuhl, J. (1993). Representation of intentions: Persisting activation in memory. Journal of Experimental Psychology: Learning, Memory, and Cognition , 19 , 1211–1226 . Goschke, T., & Kuhl, J. (1996). Remembering what to do: Explicit and implicit memory for intentions. In M. Brandimonte, G.O.Einstein, & M.A.McDaniel (Eds.), Prospective memory: Theory and applications ( pp. 53– 91 ). Mahwah, NJ: Lawrence Erlbaum Associates Inc.
116 GUYNN
Guynn, M.J. (2001). Footprints of monitoring in event-based prospective memory (Doctoral dissertation, University of New Mexico, 2001). Dissertation Abstracts International: Section B: The Sciences and Engineering , 62 , 1108 . Guynn, M.J., McDaniel, M.A., & Einstein, G.O. (2001). Remembering to perform actions: A different type of memory? In H.D.Zimmer, R.L.Cohen, M.J.Guynn, J.Engelkamp, R.Kormi-Nouri, & M.A.Foley (Eds.), Memory for action: A distinct form of episodic memory? ( pp. 25–48 ). New York: Oxford University Press. Kliegel, M, Martin, M., McDaniel, M.A., & Einstein, G. O. (2001). Varying the importance of a prospective memory task: Differential effects across time- and eventbased prospective memory. Memory , 9 , 1–11 . Lauwereyns, J. (1998). Exogenous/endogenous control of space-based/object-based attention: Four types of visual selection? European Journal of Cognitive Psychology , 10 , 41–74 . Lebiere, C, & Lee, F.J. (2001). Intention superiority effect: A context-sensitivity account. In E.M.Altmann, A. Cleeremans, C.D.Schunn, & W.D.Gray (Eds.), Proceedings of the 2001 Fourth International Conference on Cognitive Modeling ( pp. 139–144 ). Mahwah, NJ: Lawrence Erlbaum Associates Inc. Lovett, M.C, Reder, L.M., & Lebiere, C. (1999). Modeling working memory in a unified architecture: An ACT-R perspective. In A.Miyake & P.Shah (Eds.), Models of working memory: Mechanisms of active maintenance and executive control ( pp. 135–182 ). Cambridge: Cambridge University Press. Mandler, G. (1980). Recognizing: The judgment of previous occurrence. Psychological Review , 87 , 252–271 . Marsh, R.L., Hicks, J.L., & Bink, M.L. (1998). Activation of completed, uncompleted, and partially completed intentions. Journal of Experimental Psychology: Learning, Memory, and Cognition , 24 , 350–361 . Marsh, R.L., Hicks, J.L., & Bryan, E.S. (1999). The activation of unrelated and canceled intentions. Memory & Cognition , 27 , 320–327 . Marsh, R.L., Hicks, J.L., & Watson, V. (2002). The dynamics of intention retrieval and coordination of action in event-based prospective memory. Journal of Experimental Psychology: Learning, Memory, and Cognition , 28 , 652–659 . McDaniel, M.A. (1995). Prospective memory: Progress and processes. In D.L.Medin (Ed.), The psychology of learning and motivation , Vol. 33 (pp. 191–221 ). San Diego: Academic Press. McDaniel, M.A., & Einstein, G.O. (1993). The importance of cue familiarity and cue distinctiveness in prospective memory. Memory , 1 , 23–41 . McDaniel, M.A., & Einstein, G.O. (2000). Strategic and automatic processes in prospective memory retrieval: A multiprocess framework . Applied Cognitive Psychology , 14 , 127–144 . McDaniel, M.A., Guynn, M.J., Einstein, G.O., & Breneiser, J. (2003). Cue-focused and automatic- associative processes in prospective memory retrieval . Manuscript submitted for publication. McIntosh, A.R., Nyberg, L., Bookstein, F.L., & Tulving, E. (1997). Differential functional connectivity of prefrontal and medial temporal cortices during episodic memory retrieval. Human Brain Mapping , 5 , 323–327 . Meacham, J.A., & Leiman, B. (1982). Remembering to perform future actions. In U.Neisser (Ed.), Memory observed: Remembering in natural contexts ( pp. 327– 336 ). San Francisco: Freeman.
STRATEGIC MONITORING IN PROSPECTIVE MEMORY
117
Morcom, A.M., & Rugg, M.D. (2002). Getting ready to remember: The neural correlates of task set during recognition memory. NeuroReport , 13 , 149–152 . Naveh-Benjamin, M., Craik, F.I.M., Gavrilescu, D., & Anderson, N.D. (2000). Asymmetry between encoding and retrieval processes: Evidence from divided attention and a calibration analysis. Memory & Cognition , 28 , 965–976 . Nyberg, L. (1999). Functional neuroanatomy of component processes of episodic memory retrieval. In L.-G.Nilsson & H.J.Markowitsch (Eds.), Cognitive neuroscience of memory ( pp. 43–54 ). Kirkland, WA: Hogrefe & Huber. Nyberg, L., Tulving, E., Habib, R., Nilsson, L.-G., Kapur, S., Houle, S., Cabeza, R., & McIntosh, A.R. (1995). Functional brain maps of retrieval mode and recovery of episodic information. NeuroReport , 7 , 249–252 . Shallice, T., & Burgess, P.W. (1991). Deficits in strategy application following frontal lobe damage in man. Brain , 114 , 727–741 . Smith, R.E. (2000, July). Successful initiation of intentions requires capacity . Paper presented at the 1st International Conference on Prospective Memory, Hatfield, UK. Smith, R.E., & Bayen, U.J. (2002a, November). A multinomial model of event-based prospective memory . Poster presented at the 43rd annual meeting of the Psychonomic Society, Kansas City, MO. Smith, R.E., & Bayen, U.J. (2002b, November). Investigating the role of working memory capacity in prospective memory . Poster presented at the 43rd annual meeting of the Psychonomic Society, Kansas City, MO. Sternberg, S. (1969). The discovery of processing stages: Extensions of Donders’ method. Acta Psychologica , 30 , 276–315 . Toglia, M.P, & Battig, W.F. (1978). Handbook of semantic word norms . Hillsdale, NJ: Lawrence Erlbaum Associates Inc. Tulving, E. (1983). Elements of episodic memory . New York: Oxford University Press. Tulving, E. (1998). Brain/mind correlates of human memory. In M.Sabourin, F.Craik, & M.Robert (Eds.), Advances in psychological science, Vol. 2: Biological and cognitive aspects ( pp. 441–460 ). Hove, UK: Psychology Press.
118 GUYNN
Instructions to Authors
Papers must be submitted in English only to Dra. Laura Hernández-Guzmán. Authors should prepare manuscripts according to the Publication Manual of the American Psychological Association (APA) (5th ed.). All manuscripts must include at least a 300-word abstract in English. In case the manuscript is accepted for publication, assistance will be provided, if needed, for translation of the abstract into French and Spanish. When first submitting a manuscript, please, send a virus-free electronic version (Rich Text Format or Word document, with no formatting instructions except for figures and tables) of your manuscript in a single file on an attachment to:
[email protected] or one printed copy of the paper and a 3½” high density diskette containing an electronic version of the manuscript, including tables and figures, in Windows (Rich Text Format or Word document, with no formatting instructions except for figures and tables) should be submitted to the editor: Dra. Laura Hernández-Guzmán, President, The Mexican Psychological Society, Calle Indiana 260 Desp. 608, Col. Napoles, 03710, Mexico City, Mexico. The electronic version should not include complex formatting, such as Quark Xpress or Aldus PageMaker. Also the “fast save” feature must be turned off. Manuscripts should be typewritten double spaced on one side of the page only. Please leave wide margins. Instructions on preparing tables, figures, references, metrics, and abstracts all appear in the APA Publication Manual. Any manuscript which does not conform to these guidelines may be returned for necessary revision before reviewing. The Editor retains the right to reject manuscripts that do not meet established scientific or ethical standards. In case of rejection, manuscripts cannot be returned. Reviews are not conducted in a blind fashion, though it is the reviewer’s decision whether he or she is identified to the author(s) or chooses to remain anonymous. Every effort is made to have the initial review process completed less than three months after submission, but there are occasional delays due to circumstances beyond our control. Most accepted manuscripts require some revision; suggestions to facilitate the preparation of the revision will be communicated to the author(s) in the editorial decision letter. Accepted papers
120
written in incorrect English may be returned to the author for language correction; in that case, the authors are advised to contact a native speaker. The time between initial submission and appearance in print seldom exceeds 12–14 months. When first submitting a manuscript, send only copies of the figures, not original figures. With the final revision of an accepted manuscript, send glossy prints of figures, writing lightly on the back of each figure the name of the author and the number of the figure. Manuscripts should not be submitted for concurrent consideration in other journals. Duplicate publication—that is, the publication of a manuscript that has already been published wholly (in the same or another language) or in substantial part in another journal—is also prohibited. It is the responsibility of the author to ensure that the paper contains nothing that is libellous or infringes copyright. Any change of address after submission must be sent promptly to the Editor’s or an Associate Editor’s office. If available, e-mail and fax numbers should be included in the covering letter. Page proofs will be sent to the first author (unless otherwise indicated). Corrections, other than typesetter’s errors, may be charged to the author(s). Authors will receive 50 offprints of the printed article free of charge. Additional offprints are available, at a cost, and must be ordered when returning the proofs. Manuscripts should be submitted in quadruplicate. All copies should be clear, readable, and on paper of good quality and standard size. A dot matrix or unusual typeface is only acceptable if the text is clear and legible. Authors should keep at least one copy of the manuscript to guard against loss. Manuscripts for publication in the International Platform should be sent to either of the Associate Editors: Pierre L.-J. Ritchie, School of Psychology, University of Ottawa, 145 Jean-Jacques Lussier St, PO Box 450, Stn A, Ottawa, K1N 6N5, Canada, or Merry Bullock, Science Directorate, APA 750 First Street NE, Washington DC 20002, USA
